Merge branch 'master' of ssh://bitbucket.org/czan/honours

35 files changed, 21699 insertions, 0 deletions

diff --git a/clang/lib/Headers/CMakeLists.txt b/clang/lib/Headers/CMakeLists.txt
new file mode 100644
index 0000000..78141a3
--- /dev/null
+++ b/clang/lib/Headers/CMakeLists.txt
@@ -0,0 +1,90 @@
+set(files
+  altivec.h
+  avxintrin.h
+  avx2intrin.h
+  bmiintrin.h
+  bmi2intrin.h
+  emmintrin.h
+  float.h
+  fma4intrin.h
+  immintrin.h
+  iso646.h
+  limits.h
+  lzcntintrin.h
+  mm3dnow.h
+  mmintrin.h
+  mm_malloc.h
+  nmmintrin.h
+  pmmintrin.h
+  popcntintrin.h
+  smmintrin.h
+  stdalign.h
+  stdarg.h
+  stdbool.h
+  stddef.h
+  stdint.h
+  tgmath.h
+  tmmintrin.h
+  varargs.h
+  wmmintrin.h
+  x86intrin.h
+  xmmintrin.h
+  cpuid.h
+  unwind.h
+  module.map
+  )
+
+set(output_dir ${LLVM_BINARY_DIR}/lib/clang/${CLANG_VERSION}/include)
+
+# If we are in an IDE that has a configuration directory, we need to
+# create a second copy of the headers so that 'clang' can find them if
+# it's run from the build directory.
+if(MSVC_IDE OR XCODE)
+   set(other_output_dir ${LLVM_BINARY_DIR}/bin/lib/clang/${CLANG_VERSION}/include)
+endif()
+
+# Generate arm_neon.h
+clang_tablegen(arm_neon.h.inc -gen-arm-neon
+  SOURCE ${CLANG_SOURCE_DIR}/include/clang/Basic/arm_neon.td)
+
+set(out_files)
+foreach( f ${files} )
+  set( src ${CMAKE_CURRENT_SOURCE_DIR}/${f} )
+  set( dst ${output_dir}/${f} )
+  add_custom_command(OUTPUT ${dst}
+    DEPENDS ${src}
+    COMMAND ${CMAKE_COMMAND} -E copy_if_different ${src} ${dst}
+    COMMENT "Copying clang's ${f}...")
+  list(APPEND out_files ${dst})
+
+  if(other_output_dir)
+   set(other_dst ${other_output_dir}/${f})
+    add_custom_command(OUTPUT ${other_dst}
+      DEPENDS ${src}
+      COMMAND ${CMAKE_COMMAND} -E copy_if_different ${src} ${other_dst}
+      COMMENT "Copying clang's ${f}...")    
+    list(APPEND out_files ${other_dst})
+  endif()
+endforeach( f )
+
+add_custom_command(OUTPUT ${output_dir}/arm_neon.h 
+  DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/arm_neon.h.inc
+  COMMAND ${CMAKE_COMMAND} -E copy_if_different ${CMAKE_CURRENT_BINARY_DIR}/arm_neon.h.inc ${output_dir}/arm_neon.h
+  COMMENT "Copying clang's arm_neon.h...")
+list(APPEND out_files ${output_dir}/arm_neon.h)
+
+if (other_output_dir)
+    set(other_dst ${other_output_dir}/arm_neon.h)
+    add_custom_command(OUTPUT ${other_dst}
+      DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/arm_neon.h.inc
+      COMMAND ${CMAKE_COMMAND} -E copy_if_different ${CMAKE_CURRENT_BINARY_DIR}/arm_neon.h.inc ${other_dst}
+      COMMENT "Copying clang's arm_neon.h...")
+    list(APPEND out_files ${other_dst})
+endif ()
+
+add_custom_target(clang-headers ALL DEPENDS ${out_files})
+set_target_properties(clang-headers PROPERTIES FOLDER "Misc")
+
+install(FILES ${files} ${output_dir}/arm_neon.h
+  PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
+  DESTINATION lib${LLVM_LIBDIR_SUFFIX}/clang/${CLANG_VERSION}/include)
diff --git a/clang/lib/Headers/Makefile b/clang/lib/Headers/Makefile
new file mode 100644
index 0000000..42219c4
--- /dev/null
+++ b/clang/lib/Headers/Makefile
@@ -0,0 +1,64 @@
+##===- clang/lib/Headers/Makefile --------------------------*- Makefile -*-===##
+#
+#                     The LLVM Compiler Infrastructure
+#
+# This file is distributed under the University of Illinois Open Source
+# License. See LICENSE.TXT for details.
+# 
+##===----------------------------------------------------------------------===##
+
+CLANG_LEVEL := ../..
+
+BUILT_SOURCES = arm_neon.h.inc
+TABLEGEN_INC_FILES_COMMON = 1
+
+include $(CLANG_LEVEL)/Makefile
+
+CLANG_VERSION := $(word 3,$(shell grep "CLANG_VERSION " \
+	$(PROJ_OBJ_DIR)/$(CLANG_LEVEL)/include/clang/Basic/Version.inc))
+
+HeaderDir := $(PROJ_OBJ_ROOT)/$(BuildMode)/lib/clang/$(CLANG_VERSION)/include
+
+HEADERS := $(notdir $(wildcard $(PROJ_SRC_DIR)/*.h))
+
+OBJHEADERS := $(addprefix $(HeaderDir)/, $(HEADERS))
+
+
+$(OBJHEADERS): $(HeaderDir)/%.h: $(PROJ_SRC_DIR)/%.h $(HeaderDir)/.dir $(HeaderDir)/arm_neon.h
+	$(Verb) cp $< $@
+	$(Echo) Copying $(notdir $<) to build dir
+
+$(HeaderDir)/arm_neon.h: $(BUILT_SOURCES) $(HeaderDir)/.dir
+	$(Verb) cp $< $@
+	$(Echo) Copying $(notdir $<) to build dir
+
+$(HeaderDir)/module.map: $(PROJ_SRC_DIR)/module.map $(HeaderDir)/.dir
+	$(Verb) cp $< $@
+	$(Echo) Copying $(notdir $<) to build dir
+
+
+# Hook into the standard Makefile rules.
+all-local:: $(OBJHEADERS) $(HeaderDir)/module.map
+
+PROJ_headers := $(DESTDIR)$(PROJ_prefix)/lib/clang/$(CLANG_VERSION)/include
+
+INSTHEADERS := $(addprefix $(PROJ_headers)/, $(HEADERS))
+INSTHEADERS += $(PROJ_headers)/arm_neon.h
+
+$(PROJ_headers):
+	$(Verb) $(MKDIR) $@
+
+$(INSTHEADERS): $(PROJ_headers)/%.h: $(HeaderDir)/%.h | $(PROJ_headers)
+	$(Verb) $(DataInstall) $< $(PROJ_headers)
+	$(Echo) Installing compiler include file: $(notdir $<)
+
+$(PROJ_headers)/module.map: $(HeaderDir)/module.map | $(PROJ_headers)
+	$(Verb) $(DataInstall) $< $(PROJ_headers)
+	$(Echo) Installing compiler module map file: $(notdir $<)
+
+
+install-local:: $(INSTHEADERS) $(PROJ_headers)/module.map
+
+$(ObjDir)/arm_neon.h.inc.tmp : $(CLANG_LEVEL)/include/clang/Basic/arm_neon.td $(CLANG_TBLGEN) $(ObjDir)/.dir
+	$(Echo) "Building Clang arm_neon.h.inc with tblgen"
+	$(Verb) $(ClangTableGen) -gen-arm-neon -o $(call SYSPATH, $@) $<
diff --git a/clang/lib/Headers/altivec.h b/clang/lib/Headers/altivec.h
new file mode 100644
index 0000000..a225378
--- /dev/null
+++ b/clang/lib/Headers/altivec.h
@@ -0,0 +1,11856 @@
+/*===---- altivec.h - Standard header for type generic math ---------------===*\
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __ALTIVEC_H
+#define __ALTIVEC_H
+
+#ifndef __ALTIVEC__
+#error "AltiVec support not enabled"
+#endif
+
+/* constants for mapping CR6 bits to predicate result. */
+
+#define __CR6_EQ     0
+#define __CR6_EQ_REV 1
+#define __CR6_LT     2
+#define __CR6_LT_REV 3
+
+#define __ATTRS_o_ai __attribute__((__overloadable__, __always_inline__))
+
+static vector signed char __ATTRS_o_ai
+vec_perm(vector signed char a, vector signed char b, vector unsigned char c);
+
+static vector unsigned char __ATTRS_o_ai
+vec_perm(vector unsigned char a,
+         vector unsigned char b, 
+         vector unsigned char c);
+
+static vector bool char __ATTRS_o_ai
+vec_perm(vector bool char a, vector bool char b, vector unsigned char c);
+
+static vector short __ATTRS_o_ai
+vec_perm(vector short a, vector short b, vector unsigned char c);
+
+static vector unsigned short __ATTRS_o_ai
+vec_perm(vector unsigned short a,
+         vector unsigned short b, 
+         vector unsigned char c);
+
+static vector bool short __ATTRS_o_ai
+vec_perm(vector bool short a, vector bool short b, vector unsigned char c);
+
+static vector pixel __ATTRS_o_ai
+vec_perm(vector pixel a, vector pixel b, vector unsigned char c);
+
+static vector int __ATTRS_o_ai
+vec_perm(vector int a, vector int b, vector unsigned char c);
+
+static vector unsigned int __ATTRS_o_ai
+vec_perm(vector unsigned int a, vector unsigned int b, vector unsigned char c);
+
+static vector bool int __ATTRS_o_ai
+vec_perm(vector bool int a, vector bool int b, vector unsigned char c);
+
+static vector float __ATTRS_o_ai
+vec_perm(vector float a, vector float b, vector unsigned char c);
+
+/* vec_abs */
+
+#define __builtin_altivec_abs_v16qi vec_abs
+#define __builtin_altivec_abs_v8hi  vec_abs
+#define __builtin_altivec_abs_v4si  vec_abs
+
+static vector signed char __ATTRS_o_ai
+vec_abs(vector signed char a)
+{
+  return __builtin_altivec_vmaxsb(a, -a);
+}
+
+static vector signed short __ATTRS_o_ai
+vec_abs(vector signed short a)
+{
+  return __builtin_altivec_vmaxsh(a, -a);
+}
+
+static vector signed int __ATTRS_o_ai
+vec_abs(vector signed int a)
+{
+  return __builtin_altivec_vmaxsw(a, -a);
+}
+
+static vector float __ATTRS_o_ai
+vec_abs(vector float a)
+{
+  vector unsigned int res = (vector unsigned int)a 
+                            & (vector unsigned int)(0x7FFFFFFF);
+  return (vector float)res;
+}
+
+/* vec_abss */
+
+#define __builtin_altivec_abss_v16qi vec_abss
+#define __builtin_altivec_abss_v8hi  vec_abss
+#define __builtin_altivec_abss_v4si  vec_abss
+
+static vector signed char __ATTRS_o_ai
+vec_abss(vector signed char a)
+{
+  return __builtin_altivec_vmaxsb
+           (a, __builtin_altivec_vsubsbs((vector signed char)(0), a));
+}
+
+static vector signed short __ATTRS_o_ai
+vec_abss(vector signed short a)
+{
+  return __builtin_altivec_vmaxsh
+           (a, __builtin_altivec_vsubshs((vector signed short)(0), a));
+}
+
+static vector signed int __ATTRS_o_ai
+vec_abss(vector signed int a)
+{
+  return __builtin_altivec_vmaxsw
+           (a, __builtin_altivec_vsubsws((vector signed int)(0), a));
+}
+
+/* vec_add */
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector signed char a, vector signed char b)
+{
+  return a + b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a + b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector signed char a, vector bool char b)
+{
+  return a + (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector unsigned char a, vector unsigned char b)
+{
+  return a + b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a + b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector unsigned char a, vector bool char b)
+{
+  return a + (vector unsigned char)b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector short a, vector short b)
+{
+  return a + b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector bool short a, vector short b)
+{
+  return (vector short)a + b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector short a, vector bool short b)
+{
+  return a + (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector unsigned short a, vector unsigned short b)
+{
+  return a + b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a + b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector unsigned short a, vector bool short b)
+{
+  return a + (vector unsigned short)b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector int a, vector int b)
+{
+  return a + b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector bool int a, vector int b)
+{
+  return (vector int)a + b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector int a, vector bool int b)
+{
+  return a + (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector unsigned int a, vector unsigned int b)
+{
+  return a + b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a + b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector unsigned int a, vector bool int b)
+{
+  return a + (vector unsigned int)b;
+}
+
+static vector float __ATTRS_o_ai
+vec_add(vector float a, vector float b)
+{
+  return a + b;
+}
+
+/* vec_vaddubm */
+
+#define __builtin_altivec_vaddubm vec_vaddubm
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector signed char a, vector signed char b)
+{
+  return a + b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a + b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector signed char a, vector bool char b)
+{
+  return a + (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector unsigned char a, vector unsigned char b)
+{
+  return a + b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a + b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector unsigned char a, vector bool char b)
+{
+  return a + (vector unsigned char)b;
+}
+
+/* vec_vadduhm */
+
+#define __builtin_altivec_vadduhm vec_vadduhm
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector short a, vector short b)
+{
+  return a + b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector bool short a, vector short b)
+{
+  return (vector short)a + b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector short a, vector bool short b)
+{
+  return a + (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector unsigned short a, vector unsigned short b)
+{
+  return a + b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a + b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector unsigned short a, vector bool short b)
+{
+  return a + (vector unsigned short)b;
+}
+
+/* vec_vadduwm */
+
+#define __builtin_altivec_vadduwm vec_vadduwm
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector int a, vector int b)
+{
+  return a + b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector bool int a, vector int b)
+{
+  return (vector int)a + b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector int a, vector bool int b)
+{
+  return a + (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector unsigned int a, vector unsigned int b)
+{
+  return a + b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a + b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector unsigned int a, vector bool int b)
+{
+  return a + (vector unsigned int)b;
+}
+
+/* vec_vaddfp */
+
+#define __builtin_altivec_vaddfp  vec_vaddfp
+
+static vector float __attribute__((__always_inline__))
+vec_vaddfp(vector float a, vector float b)
+{
+  return a + b;
+}
+
+/* vec_addc */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_addc(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vaddcuw(a, b);
+}
+
+/* vec_vaddcuw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vaddcuw(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vaddcuw(a, b);
+}
+
+/* vec_adds */
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vaddsbs(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vaddsbs((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vaddsbs(a, (vector signed char)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vaddubs(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vaddubs((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vaddubs(a, (vector unsigned char)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector short a, vector short b)
+{
+  return __builtin_altivec_vaddshs(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vaddshs((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vaddshs(a, (vector short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vadduhs(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vadduhs((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vadduhs(a, (vector unsigned short)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector int a, vector int b)
+{
+  return __builtin_altivec_vaddsws(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vaddsws((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vaddsws(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vadduws(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vadduws((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vadduws(a, (vector unsigned int)b);
+}
+
+/* vec_vaddsbs */
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vaddsbs(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vaddsbs((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vaddsbs(a, (vector signed char)b);
+}
+
+/* vec_vaddubs */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vaddubs(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vaddubs((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vaddubs(a, (vector unsigned char)b);
+}
+
+/* vec_vaddshs */
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector short a, vector short b)
+{
+  return __builtin_altivec_vaddshs(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vaddshs((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vaddshs(a, (vector short)b);
+}
+
+/* vec_vadduhs */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vadduhs(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vadduhs((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vadduhs(a, (vector unsigned short)b);
+}
+
+/* vec_vaddsws */
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector int a, vector int b)
+{
+  return __builtin_altivec_vaddsws(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vaddsws((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vaddsws(a, (vector int)b);
+}
+
+/* vec_vadduws */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vadduws(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vadduws((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vadduws(a, (vector unsigned int)b);
+}
+
+/* vec_and */
+
+#define __builtin_altivec_vand vec_and
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector signed char a, vector signed char b)
+{
+  return a & b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a & b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector signed char a, vector bool char b)
+{
+  return a & (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector unsigned char a, vector unsigned char b)
+{
+  return a & b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a & b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector unsigned char a, vector bool char b)
+{
+  return a & (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_and(vector bool char a, vector bool char b)
+{
+  return a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector short a, vector short b)
+{
+  return a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector bool short a, vector short b)
+{
+  return (vector short)a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector short a, vector bool short b)
+{
+  return a & (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector unsigned short a, vector unsigned short b)
+{
+  return a & b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a & b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector unsigned short a, vector bool short b)
+{
+  return a & (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_and(vector bool short a, vector bool short b)
+{
+  return a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector int a, vector int b)
+{
+  return a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector bool int a, vector int b)
+{
+  return (vector int)a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector int a, vector bool int b)
+{
+  return a & (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector unsigned int a, vector unsigned int b)
+{
+  return a & b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a & b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector unsigned int a, vector bool int b)
+{
+  return a & (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_and(vector bool int a, vector bool int b)
+{
+  return a & b;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_vand */
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector signed char a, vector signed char b)
+{
+  return a & b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a & b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector signed char a, vector bool char b)
+{
+  return a & (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector unsigned char a, vector unsigned char b)
+{
+  return a & b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a & b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector unsigned char a, vector bool char b)
+{
+  return a & (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vand(vector bool char a, vector bool char b)
+{
+  return a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector short a, vector short b)
+{
+  return a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector bool short a, vector short b)
+{
+  return (vector short)a & b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector short a, vector bool short b)
+{
+  return a & (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector unsigned short a, vector unsigned short b)
+{
+  return a & b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a & b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector unsigned short a, vector bool short b)
+{
+  return a & (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vand(vector bool short a, vector bool short b)
+{
+  return a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector int a, vector int b)
+{
+  return a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector bool int a, vector int b)
+{
+  return (vector int)a & b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector int a, vector bool int b)
+{
+  return a & (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector unsigned int a, vector unsigned int b)
+{
+  return a & b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a & b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector unsigned int a, vector bool int b)
+{
+  return a & (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vand(vector bool int a, vector bool int b)
+{
+  return a & b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a & (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_andc */
+
+#define __builtin_altivec_vandc vec_andc
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector signed char a, vector signed char b)
+{
+  return a & ~b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a & ~b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector signed char a, vector bool char b)
+{
+  return a & ~(vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector unsigned char a, vector unsigned char b)
+{
+  return a & ~b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a & ~b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector unsigned char a, vector bool char b)
+{
+  return a & ~(vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_andc(vector bool char a, vector bool char b)
+{
+  return a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector short a, vector short b)
+{
+  return a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector bool short a, vector short b)
+{
+  return (vector short)a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector short a, vector bool short b)
+{
+  return a & ~(vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector unsigned short a, vector unsigned short b)
+{
+  return a & ~b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a & ~b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector unsigned short a, vector bool short b)
+{
+  return a & ~(vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_andc(vector bool short a, vector bool short b)
+{
+  return a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector int a, vector int b)
+{
+  return a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector bool int a, vector int b)
+{
+  return (vector int)a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector int a, vector bool int b)
+{
+  return a & ~(vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector unsigned int a, vector unsigned int b)
+{
+  return a & ~b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a & ~b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector unsigned int a, vector bool int b)
+{
+  return a & ~(vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_andc(vector bool int a, vector bool int b)
+{
+  return a & ~b;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_vandc */
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector signed char a, vector signed char b)
+{
+  return a & ~b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a & ~b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector signed char a, vector bool char b)
+{
+  return a & ~(vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector unsigned char a, vector unsigned char b)
+{
+  return a & ~b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a & ~b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector unsigned char a, vector bool char b)
+{
+  return a & ~(vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vandc(vector bool char a, vector bool char b)
+{
+  return a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector short a, vector short b)
+{
+  return a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector bool short a, vector short b)
+{
+  return (vector short)a & ~b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector short a, vector bool short b)
+{
+  return a & ~(vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector unsigned short a, vector unsigned short b)
+{
+  return a & ~b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a & ~b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector unsigned short a, vector bool short b)
+{
+  return a & ~(vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vandc(vector bool short a, vector bool short b)
+{
+  return a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector int a, vector int b)
+{
+  return a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector bool int a, vector int b)
+{
+  return (vector int)a & ~b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector int a, vector bool int b)
+{
+  return a & ~(vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector unsigned int a, vector unsigned int b)
+{
+  return a & ~b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a & ~b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector unsigned int a, vector bool int b)
+{
+  return a & ~(vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vandc(vector bool int a, vector bool int b)
+{
+  return a & ~b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a & ~(vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_avg */
+
+static vector signed char __ATTRS_o_ai
+vec_avg(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vavgsb(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_avg(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vavgub(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_avg(vector short a, vector short b)
+{
+  return __builtin_altivec_vavgsh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_avg(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vavguh(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_avg(vector int a, vector int b)
+{
+  return __builtin_altivec_vavgsw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_avg(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vavguw(a, b);
+}
+
+/* vec_vavgsb */
+
+static vector signed char __attribute__((__always_inline__))
+vec_vavgsb(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vavgsb(a, b);
+}
+
+/* vec_vavgub */
+
+static vector unsigned char __attribute__((__always_inline__))
+vec_vavgub(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vavgub(a, b);
+}
+
+/* vec_vavgsh */
+
+static vector short __attribute__((__always_inline__))
+vec_vavgsh(vector short a, vector short b)
+{
+  return __builtin_altivec_vavgsh(a, b);
+}
+
+/* vec_vavguh */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vavguh(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vavguh(a, b);
+}
+
+/* vec_vavgsw */
+
+static vector int __attribute__((__always_inline__))
+vec_vavgsw(vector int a, vector int b)
+{
+  return __builtin_altivec_vavgsw(a, b);
+}
+
+/* vec_vavguw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vavguw(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vavguw(a, b);
+}
+
+/* vec_ceil */
+
+static vector float __attribute__((__always_inline__))
+vec_ceil(vector float a)
+{
+  return __builtin_altivec_vrfip(a);
+}
+
+/* vec_vrfip */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfip(vector float a)
+{
+  return __builtin_altivec_vrfip(a);
+}
+
+/* vec_cmpb */
+
+static vector int __attribute__((__always_inline__))
+vec_cmpb(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpbfp(a, b);
+}
+
+/* vec_vcmpbfp */
+
+static vector int __attribute__((__always_inline__))
+vec_vcmpbfp(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpbfp(a, b);
+}
+
+/* vec_cmpeq */
+
+static vector bool char __ATTRS_o_ai
+vec_cmpeq(vector signed char a, vector signed char b)
+{
+  return (vector bool char)
+    __builtin_altivec_vcmpequb((vector char)a, (vector char)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmpeq(vector unsigned char a, vector unsigned char b)
+{
+  return (vector bool char)
+    __builtin_altivec_vcmpequb((vector char)a, (vector char)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpeq(vector short a, vector short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpequh(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpeq(vector unsigned short a, vector unsigned short b)
+{
+  return (vector bool short)
+    __builtin_altivec_vcmpequh((vector short)a, (vector short)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector int a, vector int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpequw(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector unsigned int a, vector unsigned int b)
+{
+  return (vector bool int)
+    __builtin_altivec_vcmpequw((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpeqfp(a, b);
+}
+
+/* vec_cmpge */
+
+static vector bool int __attribute__((__always_inline__))
+vec_cmpge(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(a, b);
+}
+
+/* vec_vcmpgefp */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgefp(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(a, b);
+}
+
+/* vec_cmpgt */
+
+static vector bool char __ATTRS_o_ai
+vec_cmpgt(vector signed char a, vector signed char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(a, b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmpgt(vector unsigned char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpgt(vector short a, vector short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpgt(vector unsigned short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector int a, vector int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector unsigned int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(a, b);
+}
+
+/* vec_vcmpgtsb */
+
+static vector bool char __attribute__((__always_inline__))
+vec_vcmpgtsb(vector signed char a, vector signed char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(a, b);
+}
+
+/* vec_vcmpgtub */
+
+static vector bool char __attribute__((__always_inline__))
+vec_vcmpgtub(vector unsigned char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(a, b);
+}
+
+/* vec_vcmpgtsh */
+
+static vector bool short __attribute__((__always_inline__))
+vec_vcmpgtsh(vector short a, vector short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(a, b);
+}
+
+/* vec_vcmpgtuh */
+
+static vector bool short __attribute__((__always_inline__))
+vec_vcmpgtuh(vector unsigned short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(a, b);
+}
+
+/* vec_vcmpgtsw */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtsw(vector int a, vector int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(a, b);
+}
+
+/* vec_vcmpgtuw */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtuw(vector unsigned int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(a, b);
+}
+
+/* vec_vcmpgtfp */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtfp(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(a, b);
+}
+
+/* vec_cmple */
+
+static vector bool int __attribute__((__always_inline__))
+vec_cmple(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(b, a);
+}
+
+/* vec_cmplt */
+
+static vector bool char __ATTRS_o_ai
+vec_cmplt(vector signed char a, vector signed char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(b, a);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmplt(vector unsigned char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(b, a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmplt(vector short a, vector short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(b, a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmplt(vector unsigned short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(b, a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector int a, vector int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(b, a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector unsigned int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(b, a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector float a, vector float b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(b, a);
+}
+
+/* vec_ctf */
+
+static vector float __ATTRS_o_ai
+vec_ctf(vector int a, int b)
+{
+  return __builtin_altivec_vcfsx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ctf(vector unsigned int a, int b)
+{
+  return __builtin_altivec_vcfux((vector int)a, b);
+}
+
+/* vec_vcfsx */
+
+static vector float __attribute__((__always_inline__))
+vec_vcfsx(vector int a, int b)
+{
+  return __builtin_altivec_vcfsx(a, b);
+}
+
+/* vec_vcfux */
+
+static vector float __attribute__((__always_inline__))
+vec_vcfux(vector unsigned int a, int b)
+{
+  return __builtin_altivec_vcfux((vector int)a, b);
+}
+
+/* vec_cts */
+
+static vector int __attribute__((__always_inline__))
+vec_cts(vector float a, int b)
+{
+  return __builtin_altivec_vctsxs(a, b);
+}
+
+/* vec_vctsxs */
+
+static vector int __attribute__((__always_inline__))
+vec_vctsxs(vector float a, int b)
+{
+  return __builtin_altivec_vctsxs(a, b);
+}
+
+/* vec_ctu */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_ctu(vector float a, int b)
+{
+  return __builtin_altivec_vctuxs(a, b);
+}
+
+/* vec_vctuxs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vctuxs(vector float a, int b)
+{
+  return __builtin_altivec_vctuxs(a, b);
+}
+
+/* vec_dss */
+
+static void __attribute__((__always_inline__))
+vec_dss(int a)
+{
+  __builtin_altivec_dss(a);
+}
+
+/* vec_dssall */
+
+static void __attribute__((__always_inline__))
+vec_dssall(void)
+{
+  __builtin_altivec_dssall();
+}
+
+/* vec_dst */
+
+static void __attribute__((__always_inline__))
+vec_dst(const void *a, int b, int c)
+{
+  __builtin_altivec_dst(a, b, c);
+}
+
+/* vec_dstst */
+
+static void __attribute__((__always_inline__))
+vec_dstst(const void *a, int b, int c)
+{
+  __builtin_altivec_dstst(a, b, c);
+}
+
+/* vec_dststt */
+
+static void __attribute__((__always_inline__))
+vec_dststt(const void *a, int b, int c)
+{
+  __builtin_altivec_dststt(a, b, c);
+}
+
+/* vec_dstt */
+
+static void __attribute__((__always_inline__))
+vec_dstt(const void *a, int b, int c)
+{
+  __builtin_altivec_dstt(a, b, c);
+}
+
+/* vec_expte */
+
+static vector float __attribute__((__always_inline__))
+vec_expte(vector float a)
+{
+  return __builtin_altivec_vexptefp(a);
+}
+
+/* vec_vexptefp */
+
+static vector float __attribute__((__always_inline__))
+vec_vexptefp(vector float a)
+{
+  return __builtin_altivec_vexptefp(a);
+}
+
+/* vec_floor */
+
+static vector float __attribute__((__always_inline__))
+vec_floor(vector float a)
+{
+  return __builtin_altivec_vrfim(a);
+}
+
+/* vec_vrfim */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfim(vector float a)
+{
+  return __builtin_altivec_vrfim(a);
+}
+
+/* vec_ld */
+
+static vector signed char __ATTRS_o_ai
+vec_ld(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvx(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_ld(int a, const signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ld(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ld(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_ld(int a, const vector bool char *b)
+{
+  return (vector bool char)__builtin_altivec_lvx(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ld(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvx(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ld(int a, const short *b)
+{
+  return (vector short)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ld(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ld(int a, const unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_ld(int a, const vector bool short *b)
+{
+  return (vector bool short)__builtin_altivec_lvx(a, b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_ld(int a, const vector pixel *b)
+{
+  return (vector pixel)__builtin_altivec_lvx(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ld(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvx(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ld(int a, const int *b)
+{
+  return (vector int)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ld(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ld(int a, const unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_ld(int a, const vector bool int *b)
+{
+  return (vector bool int)__builtin_altivec_lvx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ld(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ld(int a, const float *b)
+{
+  return (vector float)__builtin_altivec_lvx(a, b);
+}
+
+/* vec_lvx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvx(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvx(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvx(int a, const signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvx(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvx(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvx(int a, const vector bool char *b)
+{
+  return (vector bool char)__builtin_altivec_lvx(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvx(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvx(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvx(int a, const short *b)
+{
+  return (vector short)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvx(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvx(int a, const unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvx(int a, const vector bool short *b)
+{
+  return (vector bool short)__builtin_altivec_lvx(a, b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvx(int a, const vector pixel *b)
+{
+  return (vector pixel)__builtin_altivec_lvx(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvx(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvx(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvx(int a, const int *b)
+{
+  return (vector int)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvx(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvx(int a, const unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvx(int a, const vector bool int *b)
+{
+  return (vector bool int)__builtin_altivec_lvx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvx(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvx(int a, const float *b)
+{
+  return (vector float)__builtin_altivec_lvx(a, b);
+}
+
+/* vec_lde */
+
+static vector signed char __ATTRS_o_ai
+vec_lde(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvebx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lde(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvebx(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lde(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvehx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lde(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvehx(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lde(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvewx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lde(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvewx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lde(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvewx(a, b);
+}
+
+/* vec_lvebx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvebx(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvebx(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvebx(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvebx(a, b);
+}
+
+/* vec_lvehx */
+
+static vector short __ATTRS_o_ai
+vec_lvehx(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvehx(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvehx(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvehx(a, b);
+}
+
+/* vec_lvewx */
+
+static vector int __ATTRS_o_ai
+vec_lvewx(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvewx(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvewx(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvewx(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvewx(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvewx(a, b);
+}
+
+/* vec_ldl */
+
+static vector signed char __ATTRS_o_ai
+vec_ldl(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_ldl(int a, const signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ldl(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ldl(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_ldl(int a, const vector bool char *b)
+{
+  return (vector bool char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ldl(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ldl(int a, const short *b)
+{
+  return (vector short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ldl(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ldl(int a, const unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_ldl(int a, const vector bool short *b)
+{
+  return (vector bool short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_ldl(int a, const vector pixel *b)
+{
+  return (vector pixel short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ldl(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ldl(int a, const int *b)
+{
+  return (vector int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ldl(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ldl(int a, const unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_ldl(int a, const vector bool int *b)
+{
+  return (vector bool int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ldl(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvxl(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ldl(int a, const float *b)
+{
+  return (vector float)__builtin_altivec_lvxl(a, b);
+}
+
+/* vec_lvxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvxl(int a, const vector signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvxl(int a, const signed char *b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvxl(int a, const vector unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvxl(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvxl(int a, const vector bool char *b)
+{
+  return (vector bool char)__builtin_altivec_lvxl(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvxl(int a, const vector short *b)
+{
+  return (vector short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvxl(int a, const short *b)
+{
+  return (vector short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvxl(int a, const vector unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvxl(int a, const unsigned short *b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvxl(int a, const vector bool short *b)
+{
+  return (vector bool short)__builtin_altivec_lvxl(a, b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvxl(int a, const vector pixel *b)
+{
+  return (vector pixel)__builtin_altivec_lvxl(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvxl(int a, const vector int *b)
+{
+  return (vector int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvxl(int a, const int *b)
+{
+  return (vector int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvxl(int a, const vector unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvxl(int a, const unsigned int *b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvxl(int a, const vector bool int *b)
+{
+  return (vector bool int)__builtin_altivec_lvxl(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvxl(int a, const vector float *b)
+{
+  return (vector float)__builtin_altivec_lvxl(a, b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvxl(int a, const float *b)
+{
+  return (vector float)__builtin_altivec_lvxl(a, b);
+}
+
+/* vec_loge */
+
+static vector float __attribute__((__always_inline__))
+vec_loge(vector float a)
+{
+  return __builtin_altivec_vlogefp(a);
+}
+
+/* vec_vlogefp */
+
+static vector float __attribute__((__always_inline__))
+vec_vlogefp(vector float a)
+{
+  return __builtin_altivec_vlogefp(a);
+}
+
+/* vec_lvsl */
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const signed char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const short *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const unsigned short *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const int *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const unsigned int *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int a, const float *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(a, b);
+}
+
+/* vec_lvsr */
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const signed char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const unsigned char *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const short *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const unsigned short *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const int *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const unsigned int *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int a, const float *b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(a, b);
+}
+
+/* vec_madd */
+
+static vector float __attribute__((__always_inline__))
+vec_madd(vector float a, vector float b, vector float c)
+{
+  return __builtin_altivec_vmaddfp(a, b, c);
+}
+
+/* vec_vmaddfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vmaddfp(vector float a, vector float b, vector float c)
+{
+  return __builtin_altivec_vmaddfp(a, b, c);
+}
+
+/* vec_madds */
+
+static vector signed short __attribute__((__always_inline__))
+vec_madds(vector signed short a, vector signed short b, vector signed short c)
+{
+  return __builtin_altivec_vmhaddshs(a, b, c);
+}
+
+/* vec_vmhaddshs */
+static vector signed short __attribute__((__always_inline__))
+vec_vmhaddshs(vector signed short a,
+              vector signed short b, 
+              vector signed short c)
+{
+  return __builtin_altivec_vmhaddshs(a, b, c);
+}
+
+/* vec_max */
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmaxsb(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vmaxsb((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vmaxsb(a, (vector signed char)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmaxub(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmaxub((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vmaxub(a, (vector unsigned char)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector short a, vector short b)
+{
+  return __builtin_altivec_vmaxsh(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vmaxsh((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vmaxsh(a, (vector short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmaxuh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmaxuh((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vmaxuh(a, (vector unsigned short)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector int a, vector int b)
+{
+  return __builtin_altivec_vmaxsw(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vmaxsw((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vmaxsw(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vmaxuw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vmaxuw((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vmaxuw(a, (vector unsigned int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_max(vector float a, vector float b)
+{
+  return __builtin_altivec_vmaxfp(a, b);
+}
+
+/* vec_vmaxsb */
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmaxsb(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vmaxsb((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vmaxsb(a, (vector signed char)b);
+}
+
+/* vec_vmaxub */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmaxub(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmaxub((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vmaxub(a, (vector unsigned char)b);
+}
+
+/* vec_vmaxsh */
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector short a, vector short b)
+{
+  return __builtin_altivec_vmaxsh(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vmaxsh((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vmaxsh(a, (vector short)b);
+}
+
+/* vec_vmaxuh */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmaxuh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmaxuh((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vmaxuh(a, (vector unsigned short)b);
+}
+
+/* vec_vmaxsw */
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector int a, vector int b)
+{
+  return __builtin_altivec_vmaxsw(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vmaxsw((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vmaxsw(a, (vector int)b);
+}
+
+/* vec_vmaxuw */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vmaxuw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vmaxuw((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vmaxuw(a, (vector unsigned int)b);
+}
+
+/* vec_vmaxfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vmaxfp(vector float a, vector float b)
+{
+  return __builtin_altivec_vmaxfp(a, b);
+}
+
+/* vec_mergeh */
+
+static vector signed char __ATTRS_o_ai
+vec_mergeh(vector signed char a, vector signed char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_mergeh(vector unsigned char a, vector unsigned char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_mergeh(vector bool char a, vector bool char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector short __ATTRS_o_ai
+vec_mergeh(vector short a, vector short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mergeh(vector unsigned short a, vector unsigned short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_mergeh(vector bool short a, vector bool short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_mergeh(vector pixel a, vector pixel b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector int __ATTRS_o_ai
+vec_mergeh(vector int a, vector int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mergeh(vector unsigned int a, vector unsigned int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_mergeh(vector bool int a, vector bool int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector float __ATTRS_o_ai
+vec_mergeh(vector float a, vector float b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+/* vec_vmrghb */
+
+#define __builtin_altivec_vmrghb vec_vmrghb
+
+static vector signed char __ATTRS_o_ai
+vec_vmrghb(vector signed char a, vector signed char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmrghb(vector unsigned char a, vector unsigned char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vmrghb(vector bool char a, vector bool char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+/* vec_vmrghh */
+
+#define __builtin_altivec_vmrghh vec_vmrghh
+
+static vector short __ATTRS_o_ai
+vec_vmrghh(vector short a, vector short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmrghh(vector unsigned short a, vector unsigned short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vmrghh(vector bool short a, vector bool short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vmrghh(vector pixel a, vector pixel b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+/* vec_vmrghw */
+
+#define __builtin_altivec_vmrghw vec_vmrghw
+
+static vector int __ATTRS_o_ai
+vec_vmrghw(vector int a, vector int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmrghw(vector unsigned int a, vector unsigned int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vmrghw(vector bool int a, vector bool int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector float __ATTRS_o_ai
+vec_vmrghw(vector float a, vector float b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+/* vec_mergel */
+
+static vector signed char __ATTRS_o_ai
+vec_mergel(vector signed char a, vector signed char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_mergel(vector unsigned char a, vector unsigned char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_mergel(vector bool char a, vector bool char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector short __ATTRS_o_ai
+vec_mergel(vector short a, vector short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mergel(vector unsigned short a, vector unsigned short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_mergel(vector bool short a, vector bool short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_mergel(vector pixel a, vector pixel b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector int __ATTRS_o_ai
+vec_mergel(vector int a, vector int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mergel(vector unsigned int a, vector unsigned int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_mergel(vector bool int a, vector bool int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector float __ATTRS_o_ai
+vec_mergel(vector float a, vector float b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+/* vec_vmrglb */
+
+#define __builtin_altivec_vmrglb vec_vmrglb
+
+static vector signed char __ATTRS_o_ai
+vec_vmrglb(vector signed char a, vector signed char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmrglb(vector unsigned char a, vector unsigned char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vmrglb(vector bool char a, vector bool char b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+/* vec_vmrglh */
+
+#define __builtin_altivec_vmrglh vec_vmrglh
+
+static vector short __ATTRS_o_ai
+vec_vmrglh(vector short a, vector short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmrglh(vector unsigned short a, vector unsigned short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vmrglh(vector bool short a, vector bool short b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vmrglh(vector pixel a, vector pixel b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+/* vec_vmrglw */
+
+#define __builtin_altivec_vmrglw vec_vmrglw
+
+static vector int __ATTRS_o_ai
+vec_vmrglw(vector int a, vector int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmrglw(vector unsigned int a, vector unsigned int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vmrglw(vector bool int a, vector bool int b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector float __ATTRS_o_ai
+vec_vmrglw(vector float a, vector float b)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+/* vec_mfvscr */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_mfvscr(void)
+{
+  return __builtin_altivec_mfvscr();
+}
+
+/* vec_min */
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vminsb(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vminsb((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vminsb(a, (vector signed char)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vminub(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vminub((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vminub(a, (vector unsigned char)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector short a, vector short b)
+{
+  return __builtin_altivec_vminsh(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vminsh((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vminsh(a, (vector short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vminuh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vminuh((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vminuh(a, (vector unsigned short)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector int a, vector int b)
+{
+  return __builtin_altivec_vminsw(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vminsw((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vminsw(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vminuw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vminuw((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vminuw(a, (vector unsigned int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_min(vector float a, vector float b)
+{
+  return __builtin_altivec_vminfp(a, b);
+}
+
+/* vec_vminsb */
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vminsb(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vminsb((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vminsb(a, (vector signed char)b);
+}
+
+/* vec_vminub */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vminub(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vminub((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vminub(a, (vector unsigned char)b);
+}
+
+/* vec_vminsh */
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector short a, vector short b)
+{
+  return __builtin_altivec_vminsh(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vminsh((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vminsh(a, (vector short)b);
+}
+
+/* vec_vminuh */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vminuh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vminuh((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vminuh(a, (vector unsigned short)b);
+}
+
+/* vec_vminsw */
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector int a, vector int b)
+{
+  return __builtin_altivec_vminsw(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vminsw((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vminsw(a, (vector int)b);
+}
+
+/* vec_vminuw */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vminuw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vminuw((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vminuw(a, (vector unsigned int)b);
+}
+
+/* vec_vminfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vminfp(vector float a, vector float b)
+{
+  return __builtin_altivec_vminfp(a, b);
+}
+
+/* vec_mladd */
+
+#define __builtin_altivec_vmladduhm vec_mladd
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector short a, vector short b, vector short c)
+{
+  return a * b + c;
+}
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector short a, vector unsigned short b, vector unsigned short c)
+{
+  return a * (vector short)b + (vector short)c;
+}
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector unsigned short a, vector short b, vector short c)
+{
+  return (vector short)a * b + c;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mladd(vector unsigned short a,
+          vector unsigned short b, 
+          vector unsigned short c)
+{
+  return a * b + c;
+}
+
+/* vec_vmladduhm */
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector short a, vector short b, vector short c)
+{
+  return a * b + c;
+}
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector short a, vector unsigned short b, vector unsigned short c)
+{
+  return a * (vector short)b + (vector short)c;
+}
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector unsigned short a, vector short b, vector short c)
+{
+  return (vector short)a * b + c;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmladduhm(vector unsigned short a,
+              vector unsigned short b,
+              vector unsigned short c)
+{
+  return a * b + c;
+}
+
+/* vec_mradds */
+
+static vector short __attribute__((__always_inline__))
+vec_mradds(vector short a, vector short b, vector short c)
+{
+  return __builtin_altivec_vmhraddshs(a, b, c);
+}
+
+/* vec_vmhraddshs */
+
+static vector short __attribute__((__always_inline__))
+vec_vmhraddshs(vector short a, vector short b, vector short c)
+{
+  return __builtin_altivec_vmhraddshs(a, b, c);
+}
+
+/* vec_msum */
+
+static vector int __ATTRS_o_ai
+vec_msum(vector signed char a, vector unsigned char b, vector int c)
+{
+  return __builtin_altivec_vmsummbm(a, b, c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msum(vector unsigned char a, vector unsigned char b, vector unsigned int c)
+{
+  return __builtin_altivec_vmsumubm(a, b, c);
+}
+
+static vector int __ATTRS_o_ai
+vec_msum(vector short a, vector short b, vector int c)
+{
+  return __builtin_altivec_vmsumshm(a, b, c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msum(vector unsigned short a,
+         vector unsigned short b,
+         vector unsigned int c)
+{
+  return __builtin_altivec_vmsumuhm(a, b, c);
+}
+
+/* vec_vmsummbm */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsummbm(vector signed char a, vector unsigned char b, vector int c)
+{
+  return __builtin_altivec_vmsummbm(a, b, c);
+}
+
+/* vec_vmsumubm */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumubm(vector unsigned char a,
+             vector unsigned char b,
+             vector unsigned int c)
+{
+  return __builtin_altivec_vmsumubm(a, b, c);
+}
+
+/* vec_vmsumshm */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsumshm(vector short a, vector short b, vector int c)
+{
+  return __builtin_altivec_vmsumshm(a, b, c);
+}
+
+/* vec_vmsumuhm */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumuhm(vector unsigned short a,
+             vector unsigned short b,
+             vector unsigned int c)
+{
+  return __builtin_altivec_vmsumuhm(a, b, c);
+}
+
+/* vec_msums */
+
+static vector int __ATTRS_o_ai
+vec_msums(vector short a, vector short b, vector int c)
+{
+  return __builtin_altivec_vmsumshs(a, b, c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msums(vector unsigned short a,
+          vector unsigned short b,
+          vector unsigned int c)
+{
+  return __builtin_altivec_vmsumuhs(a, b, c);
+}
+
+/* vec_vmsumshs */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsumshs(vector short a, vector short b, vector int c)
+{
+  return __builtin_altivec_vmsumshs(a, b, c);
+}
+
+/* vec_vmsumuhs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumuhs(vector unsigned short a,
+             vector unsigned short b,
+             vector unsigned int c)
+{
+  return __builtin_altivec_vmsumuhs(a, b, c);
+}
+
+/* vec_mtvscr */
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector signed char a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned char a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool char a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector short a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned short a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool short a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector pixel a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector int a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned int a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool int a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector float a)
+{
+  __builtin_altivec_mtvscr((vector int)a);
+}
+
+/* vec_mule */
+
+static vector short __ATTRS_o_ai
+vec_mule(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmulesb(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mule(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmuleub(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_mule(vector short a, vector short b)
+{
+  return __builtin_altivec_vmulesh(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mule(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmuleuh(a, b);
+}
+
+/* vec_vmulesb */
+
+static vector short __attribute__((__always_inline__))
+vec_vmulesb(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmulesb(a, b);
+}
+
+/* vec_vmuleub */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vmuleub(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmuleub(a, b);
+}
+
+/* vec_vmulesh */
+
+static vector int __attribute__((__always_inline__))
+vec_vmulesh(vector short a, vector short b)
+{
+  return __builtin_altivec_vmulesh(a, b);
+}
+
+/* vec_vmuleuh */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmuleuh(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmuleuh(a, b);
+}
+
+/* vec_mulo */
+
+static vector short __ATTRS_o_ai
+vec_mulo(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmulosb(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mulo(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmuloub(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_mulo(vector short a, vector short b)
+{
+  return __builtin_altivec_vmulosh(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mulo(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmulouh(a, b);
+}
+
+/* vec_vmulosb */
+
+static vector short __attribute__((__always_inline__))
+vec_vmulosb(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vmulosb(a, b);
+}
+
+/* vec_vmuloub */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vmuloub(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vmuloub(a, b);
+}
+
+/* vec_vmulosh */
+
+static vector int __attribute__((__always_inline__))
+vec_vmulosh(vector short a, vector short b)
+{
+  return __builtin_altivec_vmulosh(a, b);
+}
+
+/* vec_vmulouh */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmulouh(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vmulouh(a, b);
+}
+
+/* vec_nmsub */
+
+static vector float __attribute__((__always_inline__))
+vec_nmsub(vector float a, vector float b, vector float c)
+{
+  return __builtin_altivec_vnmsubfp(a, b, c);
+}
+
+/* vec_vnmsubfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vnmsubfp(vector float a, vector float b, vector float c)
+{
+  return __builtin_altivec_vnmsubfp(a, b, c);
+}
+
+/* vec_nor */
+
+#define __builtin_altivec_vnor vec_nor
+
+static vector signed char __ATTRS_o_ai
+vec_nor(vector signed char a, vector signed char b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_nor(vector unsigned char a, vector unsigned char b)
+{
+  return ~(a | b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_nor(vector bool char a, vector bool char b)
+{
+  return ~(a | b);
+}
+
+static vector short __ATTRS_o_ai
+vec_nor(vector short a, vector short b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_nor(vector unsigned short a, vector unsigned short b)
+{
+  return ~(a | b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_nor(vector bool short a, vector bool short b)
+{
+  return ~(a | b);
+}
+
+static vector int __ATTRS_o_ai
+vec_nor(vector int a, vector int b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_nor(vector unsigned int a, vector unsigned int b)
+{
+  return ~(a | b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_nor(vector bool int a, vector bool int b)
+{
+  return ~(a | b);
+}
+
+static vector float __ATTRS_o_ai
+vec_nor(vector float a, vector float b)
+{
+  vector unsigned int res = ~((vector unsigned int)a | (vector unsigned int)b);
+  return (vector float)res;
+}
+
+/* vec_vnor */
+
+static vector signed char __ATTRS_o_ai
+vec_vnor(vector signed char a, vector signed char b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vnor(vector unsigned char a, vector unsigned char b)
+{
+  return ~(a | b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vnor(vector bool char a, vector bool char b)
+{
+  return ~(a | b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vnor(vector short a, vector short b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vnor(vector unsigned short a, vector unsigned short b)
+{
+  return ~(a | b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vnor(vector bool short a, vector bool short b)
+{
+  return ~(a | b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vnor(vector int a, vector int b)
+{
+  return ~(a | b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vnor(vector unsigned int a, vector unsigned int b)
+{
+  return ~(a | b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vnor(vector bool int a, vector bool int b)
+{
+  return ~(a | b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vnor(vector float a, vector float b)
+{
+  vector unsigned int res = ~((vector unsigned int)a | (vector unsigned int)b);
+  return (vector float)res;
+}
+
+/* vec_or */
+
+#define __builtin_altivec_vor vec_or
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector signed char a, vector signed char b)
+{
+  return a | b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a | b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector signed char a, vector bool char b)
+{
+  return a | (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector unsigned char a, vector unsigned char b)
+{
+  return a | b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a | b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector unsigned char a, vector bool char b)
+{
+  return a | (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_or(vector bool char a, vector bool char b)
+{
+  return a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector short a, vector short b)
+{
+  return a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector bool short a, vector short b)
+{
+  return (vector short)a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector short a, vector bool short b)
+{
+  return a | (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector unsigned short a, vector unsigned short b)
+{
+  return a | b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a | b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector unsigned short a, vector bool short b)
+{
+  return a | (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_or(vector bool short a, vector bool short b)
+{
+  return a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector int a, vector int b)
+{
+  return a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector bool int a, vector int b)
+{
+  return (vector int)a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector int a, vector bool int b)
+{
+  return a | (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector unsigned int a, vector unsigned int b)
+{
+  return a | b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a | b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector unsigned int a, vector bool int b)
+{
+  return a | (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_or(vector bool int a, vector bool int b)
+{
+  return a | b;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_vor */
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector signed char a, vector signed char b)
+{
+  return a | b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a | b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector signed char a, vector bool char b)
+{
+  return a | (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector unsigned char a, vector unsigned char b)
+{
+  return a | b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a | b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector unsigned char a, vector bool char b)
+{
+  return a | (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vor(vector bool char a, vector bool char b)
+{
+  return a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector short a, vector short b)
+{
+  return a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector bool short a, vector short b)
+{
+  return (vector short)a | b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector short a, vector bool short b)
+{
+  return a | (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector unsigned short a, vector unsigned short b)
+{
+  return a | b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a | b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector unsigned short a, vector bool short b)
+{
+  return a | (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vor(vector bool short a, vector bool short b)
+{
+  return a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector int a, vector int b)
+{
+  return a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector bool int a, vector int b)
+{
+  return (vector int)a | b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector int a, vector bool int b)
+{
+  return a | (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector unsigned int a, vector unsigned int b)
+{
+  return a | b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a | b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector unsigned int a, vector bool int b)
+{
+  return a | (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vor(vector bool int a, vector bool int b)
+{
+  return a | b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a | (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_pack */
+
+static vector signed char __ATTRS_o_ai
+vec_pack(vector signed short a, vector signed short b)
+{
+  return (vector signed char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_pack(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_pack(vector bool short a, vector bool short b)
+{
+  return (vector bool char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+static vector short __ATTRS_o_ai
+vec_pack(vector int a, vector int b)
+{
+  return (vector short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_pack(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_pack(vector bool int a, vector bool int b)
+{
+  return (vector bool short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+/* vec_vpkuhum */
+
+#define __builtin_altivec_vpkuhum vec_vpkuhum
+
+static vector signed char __ATTRS_o_ai
+vec_vpkuhum(vector signed short a, vector signed short b)
+{
+  return (vector signed char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkuhum(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vpkuhum(vector bool short a, vector bool short b)
+{
+  return (vector bool char)vec_perm(a, b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+}
+
+/* vec_vpkuwum */
+
+#define __builtin_altivec_vpkuwum vec_vpkuwum
+
+static vector short __ATTRS_o_ai
+vec_vpkuwum(vector int a, vector int b)
+{
+  return (vector short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkuwum(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vpkuwum(vector bool int a, vector bool int b)
+{
+  return (vector bool short)vec_perm(a, b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+}
+
+/* vec_packpx */
+
+static vector pixel __attribute__((__always_inline__))
+vec_packpx(vector unsigned int a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vpkpx(a, b);
+}
+
+/* vec_vpkpx */
+
+static vector pixel __attribute__((__always_inline__))
+vec_vpkpx(vector unsigned int a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vpkpx(a, b);
+}
+
+/* vec_packs */
+
+static vector signed char __ATTRS_o_ai
+vec_packs(vector short a, vector short b)
+{
+  return __builtin_altivec_vpkshss(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_packs(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vpkuhus(a, b);
+}
+
+static vector signed short __ATTRS_o_ai
+vec_packs(vector int a, vector int b)
+{
+  return __builtin_altivec_vpkswss(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packs(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vpkuwus(a, b);
+}
+
+/* vec_vpkshss */
+
+static vector signed char __attribute__((__always_inline__))
+vec_vpkshss(vector short a, vector short b)
+{
+  return __builtin_altivec_vpkshss(a, b);
+}
+
+/* vec_vpkuhus */
+
+static vector unsigned char __attribute__((__always_inline__))
+vec_vpkuhus(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vpkuhus(a, b);
+}
+
+/* vec_vpkswss */
+
+static vector signed short __attribute__((__always_inline__))
+vec_vpkswss(vector int a, vector int b)
+{
+  return __builtin_altivec_vpkswss(a, b);
+}
+
+/* vec_vpkuwus */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vpkuwus(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vpkuwus(a, b);
+}
+
+/* vec_packsu */
+
+static vector unsigned char __ATTRS_o_ai
+vec_packsu(vector short a, vector short b)
+{
+  return __builtin_altivec_vpkshus(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_packsu(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vpkuhus(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packsu(vector int a, vector int b)
+{
+  return __builtin_altivec_vpkswus(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packsu(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vpkuwus(a, b);
+}
+
+/* vec_vpkshus */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkshus(vector short a, vector short b)
+{
+  return __builtin_altivec_vpkshus(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkshus(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vpkuhus(a, b);
+}
+
+/* vec_vpkswus */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkswus(vector int a, vector int b)
+{
+  return __builtin_altivec_vpkswus(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkswus(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vpkuwus(a, b);
+}
+
+/* vec_perm */
+
+vector signed char __ATTRS_o_ai
+vec_perm(vector signed char a, vector signed char b, vector unsigned char c)
+{
+  return (vector signed char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector unsigned char __ATTRS_o_ai
+vec_perm(vector unsigned char a,
+         vector unsigned char b,
+         vector unsigned char c)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool char __ATTRS_o_ai
+vec_perm(vector bool char a, vector bool char b, vector unsigned char c)
+{
+  return (vector bool char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector short __ATTRS_o_ai
+vec_perm(vector short a, vector short b, vector unsigned char c)
+{
+  return (vector short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector unsigned short __ATTRS_o_ai
+vec_perm(vector unsigned short a,
+         vector unsigned short b,
+         vector unsigned char c)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool short __ATTRS_o_ai
+vec_perm(vector bool short a, vector bool short b, vector unsigned char c)
+{
+  return (vector bool short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector pixel __ATTRS_o_ai
+vec_perm(vector pixel a, vector pixel b, vector unsigned char c)
+{
+  return (vector pixel)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector int __ATTRS_o_ai
+vec_perm(vector int a, vector int b, vector unsigned char c)
+{
+  return (vector int)__builtin_altivec_vperm_4si(a, b, c);
+}
+
+vector unsigned int __ATTRS_o_ai
+vec_perm(vector unsigned int a, vector unsigned int b, vector unsigned char c)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool int __ATTRS_o_ai
+vec_perm(vector bool int a, vector bool int b, vector unsigned char c)
+{
+  return (vector bool int)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector float __ATTRS_o_ai
+vec_perm(vector float a, vector float b, vector unsigned char c)
+{
+  return (vector float)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+/* vec_vperm */
+
+vector signed char __ATTRS_o_ai
+vec_vperm(vector signed char a, vector signed char b, vector unsigned char c)
+{
+  return (vector signed char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector unsigned char __ATTRS_o_ai
+vec_vperm(vector unsigned char a,
+          vector unsigned char b,
+          vector unsigned char c)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool char __ATTRS_o_ai
+vec_vperm(vector bool char a, vector bool char b, vector unsigned char c)
+{
+  return (vector bool char)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector short __ATTRS_o_ai
+vec_vperm(vector short a, vector short b, vector unsigned char c)
+{
+  return (vector short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector unsigned short __ATTRS_o_ai
+vec_vperm(vector unsigned short a,
+          vector unsigned short b,
+          vector unsigned char c)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool short __ATTRS_o_ai
+vec_vperm(vector bool short a, vector bool short b, vector unsigned char c)
+{
+  return (vector bool short)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector pixel __ATTRS_o_ai
+vec_vperm(vector pixel a, vector pixel b, vector unsigned char c)
+{
+  return (vector pixel)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector int __ATTRS_o_ai
+vec_vperm(vector int a, vector int b, vector unsigned char c)
+{
+  return (vector int)__builtin_altivec_vperm_4si(a, b, c);
+}
+
+vector unsigned int __ATTRS_o_ai
+vec_vperm(vector unsigned int a, vector unsigned int b, vector unsigned char c)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector bool int __ATTRS_o_ai
+vec_vperm(vector bool int a, vector bool int b, vector unsigned char c)
+{
+  return (vector bool int)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+vector float __ATTRS_o_ai
+vec_vperm(vector float a, vector float b, vector unsigned char c)
+{
+  return (vector float)
+           __builtin_altivec_vperm_4si((vector int)a, (vector int)b, c);
+}
+
+/* vec_re */
+
+vector float __attribute__((__always_inline__))
+vec_re(vector float a)
+{
+  return __builtin_altivec_vrefp(a);
+}
+
+/* vec_vrefp */
+
+vector float __attribute__((__always_inline__))
+vec_vrefp(vector float a)
+{
+  return __builtin_altivec_vrefp(a);
+}
+
+/* vec_rl */
+
+static vector signed char __ATTRS_o_ai
+vec_rl(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)__builtin_altivec_vrlb((vector char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_rl(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)__builtin_altivec_vrlb((vector char)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_rl(vector short a, vector unsigned short b)
+{
+  return __builtin_altivec_vrlh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_rl(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)__builtin_altivec_vrlh((vector short)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_rl(vector int a, vector unsigned int b)
+{
+  return __builtin_altivec_vrlw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_rl(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)__builtin_altivec_vrlw((vector int)a, b);
+}
+
+/* vec_vrlb */
+
+static vector signed char __ATTRS_o_ai
+vec_vrlb(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)__builtin_altivec_vrlb((vector char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vrlb(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)__builtin_altivec_vrlb((vector char)a, b);
+}
+
+/* vec_vrlh */
+
+static vector short __ATTRS_o_ai
+vec_vrlh(vector short a, vector unsigned short b)
+{
+  return __builtin_altivec_vrlh(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vrlh(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)__builtin_altivec_vrlh((vector short)a, b);
+}
+
+/* vec_vrlw */
+
+static vector int __ATTRS_o_ai
+vec_vrlw(vector int a, vector unsigned int b)
+{
+  return __builtin_altivec_vrlw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vrlw(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)__builtin_altivec_vrlw((vector int)a, b);
+}
+
+/* vec_round */
+
+static vector float __attribute__((__always_inline__))
+vec_round(vector float a)
+{
+  return __builtin_altivec_vrfin(a);
+}
+
+/* vec_vrfin */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfin(vector float a)
+{
+  return __builtin_altivec_vrfin(a);
+}
+
+/* vec_rsqrte */
+
+static __vector float __attribute__((__always_inline__))
+vec_rsqrte(vector float a)
+{
+  return __builtin_altivec_vrsqrtefp(a);
+}
+
+/* vec_vrsqrtefp */
+
+static __vector float __attribute__((__always_inline__))
+vec_vrsqrtefp(vector float a)
+{
+  return __builtin_altivec_vrsqrtefp(a);
+}
+
+/* vec_sel */
+
+#define __builtin_altivec_vsel_4si vec_sel
+
+static vector signed char __ATTRS_o_ai
+vec_sel(vector signed char a, vector signed char b, vector unsigned char c)
+{
+  return (a & ~(vector signed char)c) | (b & (vector signed char)c);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sel(vector signed char a, vector signed char b, vector bool char c)
+{
+  return (a & ~(vector signed char)c) | (b & (vector signed char)c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sel(vector unsigned char a, vector unsigned char b, vector unsigned char c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sel(vector unsigned char a, vector unsigned char b, vector bool char c)
+{
+  return (a & ~(vector unsigned char)c) | (b & (vector unsigned char)c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sel(vector bool char a, vector bool char b, vector unsigned char c)
+{
+  return (a & ~(vector bool char)c) | (b & (vector bool char)c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sel(vector bool char a, vector bool char b, vector bool char c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector short __ATTRS_o_ai
+vec_sel(vector short a, vector short b, vector unsigned short c)
+{
+  return (a & ~(vector short)c) | (b & (vector short)c);
+}
+
+static vector short __ATTRS_o_ai
+vec_sel(vector short a, vector short b, vector bool short c)
+{
+  return (a & ~(vector short)c) | (b & (vector short)c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sel(vector unsigned short a,
+        vector unsigned short b,
+        vector unsigned short c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sel(vector unsigned short a, vector unsigned short b, vector bool short c)
+{
+  return (a & ~(vector unsigned short)c) | (b & (vector unsigned short)c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sel(vector bool short a, vector bool short b, vector unsigned short c)
+{
+  return (a & ~(vector bool short)c) | (b & (vector bool short)c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sel(vector bool short a, vector bool short b, vector bool short c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector int __ATTRS_o_ai
+vec_sel(vector int a, vector int b, vector unsigned int c)
+{
+  return (a & ~(vector int)c) | (b & (vector int)c);
+}
+
+static vector int __ATTRS_o_ai
+vec_sel(vector int a, vector int b, vector bool int c)
+{
+  return (a & ~(vector int)c) | (b & (vector int)c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sel(vector unsigned int a, vector unsigned int b, vector unsigned int c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sel(vector unsigned int a, vector unsigned int b, vector bool int c)
+{
+  return (a & ~(vector unsigned int)c) | (b & (vector unsigned int)c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sel(vector bool int a, vector bool int b, vector unsigned int c)
+{
+  return (a & ~(vector bool int)c) | (b & (vector bool int)c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sel(vector bool int a, vector bool int b, vector bool int c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector float __ATTRS_o_ai
+vec_sel(vector float a, vector float b, vector unsigned int c)
+{
+  vector int res = ((vector int)a & ~(vector int)c) 
+                   | ((vector int)b & (vector int)c);
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_sel(vector float a, vector float b, vector bool int c)
+{
+  vector int res = ((vector int)a & ~(vector int)c)
+                   | ((vector int)b & (vector int)c);
+  return (vector float)res;
+}
+
+/* vec_vsel */
+
+static vector signed char __ATTRS_o_ai
+vec_vsel(vector signed char a, vector signed char b, vector unsigned char c)
+{
+  return (a & ~(vector signed char)c) | (b & (vector signed char)c);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsel(vector signed char a, vector signed char b, vector bool char c)
+{
+  return (a & ~(vector signed char)c) | (b & (vector signed char)c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsel(vector unsigned char a, vector unsigned char b, vector unsigned char c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsel(vector unsigned char a, vector unsigned char b, vector bool char c)
+{
+  return (a & ~(vector unsigned char)c) | (b & (vector unsigned char)c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsel(vector bool char a, vector bool char b, vector unsigned char c)
+{
+  return (a & ~(vector bool char)c) | (b & (vector bool char)c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsel(vector bool char a, vector bool char b, vector bool char c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsel(vector short a, vector short b, vector unsigned short c)
+{
+  return (a & ~(vector short)c) | (b & (vector short)c);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsel(vector short a, vector short b, vector bool short c)
+{
+  return (a & ~(vector short)c) | (b & (vector short)c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsel(vector unsigned short a,
+         vector unsigned short b,
+         vector unsigned short c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsel(vector unsigned short a, vector unsigned short b, vector bool short c)
+{
+  return (a & ~(vector unsigned short)c) | (b & (vector unsigned short)c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsel(vector bool short a, vector bool short b, vector unsigned short c)
+{
+  return (a & ~(vector bool short)c) | (b & (vector bool short)c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsel(vector bool short a, vector bool short b, vector bool short c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsel(vector int a, vector int b, vector unsigned int c)
+{
+  return (a & ~(vector int)c) | (b & (vector int)c);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsel(vector int a, vector int b, vector bool int c)
+{
+  return (a & ~(vector int)c) | (b & (vector int)c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsel(vector unsigned int a, vector unsigned int b, vector unsigned int c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsel(vector unsigned int a, vector unsigned int b, vector bool int c)
+{
+  return (a & ~(vector unsigned int)c) | (b & (vector unsigned int)c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsel(vector bool int a, vector bool int b, vector unsigned int c)
+{
+  return (a & ~(vector bool int)c) | (b & (vector bool int)c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsel(vector bool int a, vector bool int b, vector bool int c)
+{
+  return (a & ~c) | (b & c);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsel(vector float a, vector float b, vector unsigned int c)
+{
+  vector int res = ((vector int)a & ~(vector int)c)
+                   | ((vector int)b & (vector int)c);
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vsel(vector float a, vector float b, vector bool int c)
+{
+  vector int res = ((vector int)a & ~(vector int)c)
+                   | ((vector int)b & (vector int)c);
+  return (vector float)res;
+}
+
+/* vec_sl */
+
+static vector signed char __ATTRS_o_ai
+vec_sl(vector signed char a, vector unsigned char b)
+{
+  return a << (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sl(vector unsigned char a, vector unsigned char b)
+{
+  return a << b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sl(vector short a, vector unsigned short b)
+{
+  return a << (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sl(vector unsigned short a, vector unsigned short b)
+{
+  return a << b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sl(vector int a, vector unsigned int b)
+{
+  return a << (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sl(vector unsigned int a, vector unsigned int b)
+{
+  return a << b;
+}
+
+/* vec_vslb */
+
+#define __builtin_altivec_vslb vec_vslb
+
+static vector signed char __ATTRS_o_ai
+vec_vslb(vector signed char a, vector unsigned char b)
+{
+  return vec_sl(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslb(vector unsigned char a, vector unsigned char b)
+{
+  return vec_sl(a, b);
+}
+
+/* vec_vslh */
+
+#define __builtin_altivec_vslh vec_vslh
+
+static vector short __ATTRS_o_ai
+vec_vslh(vector short a, vector unsigned short b)
+{
+  return vec_sl(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslh(vector unsigned short a, vector unsigned short b)
+{
+  return vec_sl(a, b);
+}
+
+/* vec_vslw */
+
+#define __builtin_altivec_vslw vec_vslw
+
+static vector int __ATTRS_o_ai
+vec_vslw(vector int a, vector unsigned int b)
+{
+  return vec_sl(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslw(vector unsigned int a, vector unsigned int b)
+{
+  return vec_sl(a, b);
+}
+
+/* vec_sld */
+
+#define __builtin_altivec_vsldoi_4si vec_sld
+
+static vector signed char __ATTRS_o_ai
+vec_sld(vector signed char a, vector signed char b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sld(vector unsigned char a, vector unsigned char b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector short __ATTRS_o_ai
+vec_sld(vector short a, vector short b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sld(vector unsigned short a, vector unsigned short b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sld(vector pixel a, vector pixel b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector int __ATTRS_o_ai
+vec_sld(vector int a, vector int b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sld(vector unsigned int a, vector unsigned int b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector float __ATTRS_o_ai
+vec_sld(vector float a, vector float b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+/* vec_vsldoi */
+
+static vector signed char __ATTRS_o_ai
+vec_vsldoi(vector signed char a, vector signed char b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsldoi(vector unsigned char a, vector unsigned char b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector short __ATTRS_o_ai
+vec_vsldoi(vector short a, vector short b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsldoi(vector unsigned short a, vector unsigned short b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsldoi(vector pixel a, vector pixel b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector int __ATTRS_o_ai
+vec_vsldoi(vector int a, vector int b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsldoi(vector unsigned int a, vector unsigned int b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+static vector float __ATTRS_o_ai
+vec_vsldoi(vector float a, vector float b, unsigned char c)
+{
+  return vec_perm(a, b, (vector unsigned char)
+    (c,   c+1, c+2,  c+3,  c+4,  c+5,  c+6,  c+7, 
+     c+8, c+9, c+10, c+11, c+12, c+13, c+14, c+15));
+}
+
+/* vec_sll */
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char a, vector unsigned short b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char a, vector unsigned int b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char a, vector unsigned short b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char a, vector unsigned int b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char a, vector unsigned short b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char a, vector unsigned int b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short a, vector unsigned short b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short a, vector unsigned int b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short a, vector unsigned int b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short a, vector unsigned char b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short a, vector unsigned int b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel a, vector unsigned short b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int a, vector unsigned short b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int a, vector unsigned int b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int a, vector unsigned short b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int a, vector unsigned char b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int a, vector unsigned short b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+/* vec_vsl */
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char a, vector unsigned short b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char a, vector unsigned int b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char a, vector unsigned short b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char a, vector unsigned int b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char a, vector unsigned short b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char a, vector unsigned int b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short a, vector unsigned short b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short a, vector unsigned int b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short a, vector unsigned int b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short a, vector unsigned char b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short a, vector unsigned int b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel a, vector unsigned short b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int a, vector unsigned short b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int a, vector unsigned int b)
+{
+  return (vector int)__builtin_altivec_vsl(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int a, vector unsigned short b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int a, vector unsigned char b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int a, vector unsigned short b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)a, (vector int)b);
+}
+
+/* vec_slo */
+
+static vector signed char __ATTRS_o_ai
+vec_slo(vector signed char a, vector signed char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_slo(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_slo(vector unsigned char a, vector signed char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_slo(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_slo(vector short a, vector signed char b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_slo(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_slo(vector unsigned short a, vector signed char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_slo(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_slo(vector pixel a, vector signed char b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_slo(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_slo(vector int a, vector signed char b)
+{
+  return (vector int)__builtin_altivec_vslo(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_slo(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vslo(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_slo(vector unsigned int a, vector signed char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_slo(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_slo(vector float a, vector signed char b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_slo(vector float a, vector unsigned char b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+/* vec_vslo */
+
+static vector signed char __ATTRS_o_ai
+vec_vslo(vector signed char a, vector signed char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vslo(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslo(vector unsigned char a, vector signed char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslo(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vslo(vector short a, vector signed char b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vslo(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslo(vector unsigned short a, vector signed char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslo(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vslo(vector pixel a, vector signed char b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vslo(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vslo(vector int a, vector signed char b)
+{
+  return (vector int)__builtin_altivec_vslo(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vslo(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vslo(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslo(vector unsigned int a, vector signed char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslo(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vslo(vector float a, vector signed char b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vslo(vector float a, vector unsigned char b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)a, (vector int)b);
+}
+
+/* vec_splat */
+
+static vector signed char __ATTRS_o_ai
+vec_splat(vector signed char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_splat(vector unsigned char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_splat(vector bool char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+static vector short __ATTRS_o_ai
+vec_splat(vector short a, unsigned char b)
+{ 
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_splat(vector unsigned short a, unsigned char b)
+{ 
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_splat(vector bool short a, unsigned char b)
+{ 
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_splat(vector pixel a, unsigned char b)
+{ 
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector int __ATTRS_o_ai
+vec_splat(vector int a, unsigned char b)
+{ 
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_splat(vector unsigned int a, unsigned char b)
+{ 
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_splat(vector bool int a, unsigned char b)
+{ 
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector float __ATTRS_o_ai
+vec_splat(vector float a, unsigned char b)
+{ 
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+/* vec_vspltb */
+
+#define __builtin_altivec_vspltb vec_vspltb
+
+static vector signed char __ATTRS_o_ai
+vec_vspltb(vector signed char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vspltb(vector unsigned char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vspltb(vector bool char a, unsigned char b)
+{
+  return vec_perm(a, a, (vector unsigned char)(b));
+}
+
+/* vec_vsplth */
+
+#define __builtin_altivec_vsplth vec_vsplth
+
+static vector short __ATTRS_o_ai
+vec_vsplth(vector short a, unsigned char b)
+{
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsplth(vector unsigned short a, unsigned char b)
+{
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsplth(vector bool short a, unsigned char b)
+{
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsplth(vector pixel a, unsigned char b)
+{
+  b *= 2;
+  unsigned char b1=b+1;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1, b, b1));
+}
+
+/* vec_vspltw */
+
+#define __builtin_altivec_vspltw vec_vspltw
+
+static vector int __ATTRS_o_ai
+vec_vspltw(vector int a, unsigned char b)
+{
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vspltw(vector unsigned int a, unsigned char b)
+{
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vspltw(vector bool int a, unsigned char b)
+{
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+static vector float __ATTRS_o_ai
+vec_vspltw(vector float a, unsigned char b)
+{
+  b *= 4;
+  unsigned char b1=b+1, b2=b+2, b3=b+3;
+  return vec_perm(a, a, (vector unsigned char)
+    (b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3, b, b1, b2, b3));
+}
+
+/* vec_splat_s8 */
+
+#define __builtin_altivec_vspltisb vec_splat_s8
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector signed char __ATTRS_o_ai
+vec_splat_s8(signed char a)
+{
+  return (vector signed char)(a);
+}
+
+/* vec_vspltisb */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector signed char __ATTRS_o_ai
+vec_vspltisb(signed char a)
+{
+  return (vector signed char)(a);
+}
+
+/* vec_splat_s16 */
+
+#define __builtin_altivec_vspltish vec_splat_s16
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector short __ATTRS_o_ai
+vec_splat_s16(signed char a)
+{
+  return (vector short)(a);
+}
+
+/* vec_vspltish */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector short __ATTRS_o_ai
+vec_vspltish(signed char a)
+{
+  return (vector short)(a);
+}
+
+/* vec_splat_s32 */
+
+#define __builtin_altivec_vspltisw vec_splat_s32
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector int __ATTRS_o_ai
+vec_splat_s32(signed char a)
+{
+  return (vector int)(a);
+}
+
+/* vec_vspltisw */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector int __ATTRS_o_ai
+vec_vspltisw(signed char a)
+{
+  return (vector int)(a);
+}
+
+/* vec_splat_u8 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned char __ATTRS_o_ai
+vec_splat_u8(unsigned char a)
+{
+  return (vector unsigned char)(a);
+}
+
+/* vec_splat_u16 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned short __ATTRS_o_ai
+vec_splat_u16(signed char a)
+{
+  return (vector unsigned short)(a);
+}
+
+/* vec_splat_u32 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned int __ATTRS_o_ai
+vec_splat_u32(signed char a)
+{
+  return (vector unsigned int)(a);
+}
+
+/* vec_sr */
+
+static vector signed char __ATTRS_o_ai
+vec_sr(vector signed char a, vector unsigned char b)
+{
+  return a >> (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sr(vector unsigned char a, vector unsigned char b)
+{
+  return a >> b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sr(vector short a, vector unsigned short b)
+{
+  return a >> (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sr(vector unsigned short a, vector unsigned short b)
+{
+  return a >> b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sr(vector int a, vector unsigned int b)
+{
+  return a >> (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sr(vector unsigned int a, vector unsigned int b)
+{
+  return a >> b;
+}
+
+/* vec_vsrb */
+
+#define __builtin_altivec_vsrb vec_vsrb
+
+static vector signed char __ATTRS_o_ai
+vec_vsrb(vector signed char a, vector unsigned char b)
+{
+  return a >> (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsrb(vector unsigned char a, vector unsigned char b)
+{
+  return a >> b;
+}
+
+/* vec_vsrh */
+
+#define __builtin_altivec_vsrh vec_vsrh
+
+static vector short __ATTRS_o_ai
+vec_vsrh(vector short a, vector unsigned short b)
+{
+  return a >> (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsrh(vector unsigned short a, vector unsigned short b)
+{
+  return a >> b;
+}
+
+/* vec_vsrw */
+
+#define __builtin_altivec_vsrw vec_vsrw
+
+static vector int __ATTRS_o_ai
+vec_vsrw(vector int a, vector unsigned int b)
+{
+  return a >> (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsrw(vector unsigned int a, vector unsigned int b)
+{
+  return a >> b;
+}
+
+/* vec_sra */
+
+static vector signed char __ATTRS_o_ai
+vec_sra(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)__builtin_altivec_vsrab((vector char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sra(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)__builtin_altivec_vsrab((vector char)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sra(vector short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsrah(a, (vector unsigned short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sra(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)__builtin_altivec_vsrah((vector short)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sra(vector int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsraw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sra(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)__builtin_altivec_vsraw((vector int)a, b);
+}
+
+/* vec_vsrab */
+
+static vector signed char __ATTRS_o_ai
+vec_vsrab(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)__builtin_altivec_vsrab((vector char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsrab(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)__builtin_altivec_vsrab((vector char)a, b);
+}
+
+/* vec_vsrah */
+
+static vector short __ATTRS_o_ai
+vec_vsrah(vector short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsrah(a, (vector unsigned short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsrah(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)__builtin_altivec_vsrah((vector short)a, b);
+}
+
+/* vec_vsraw */
+
+static vector int __ATTRS_o_ai
+vec_vsraw(vector int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsraw(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsraw(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)__builtin_altivec_vsraw((vector int)a, b);
+}
+
+/* vec_srl */
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char a, vector unsigned short b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char a, vector unsigned int b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char a, vector unsigned short b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char a, vector unsigned int b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char a, vector unsigned short b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char a, vector unsigned int b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short a, vector unsigned short b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short a, vector unsigned int b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short a, vector unsigned int b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short a, vector unsigned char b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short a, vector unsigned int b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel a, vector unsigned short b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int a, vector unsigned short b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int a, vector unsigned int b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int a, vector unsigned short b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int a, vector unsigned char b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int a, vector unsigned short b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+/* vec_vsr */
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char a, vector unsigned short b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char a, vector unsigned int b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char a, vector unsigned short b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char a, vector unsigned int b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char a, vector unsigned char b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char a, vector unsigned short b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char a, vector unsigned int b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short a, vector unsigned short b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short a, vector unsigned int b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short a, vector unsigned short b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short a, vector unsigned int b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short a, vector unsigned char b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short a, vector unsigned short b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short a, vector unsigned int b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel a, vector unsigned short b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel a, vector unsigned int b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int a, vector unsigned short b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int a, vector unsigned int b)
+{
+  return (vector int)__builtin_altivec_vsr(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int a, vector unsigned short b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int a, vector unsigned int b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int a, vector unsigned char b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int a, vector unsigned short b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int a, vector unsigned int b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)a, (vector int)b);
+}
+
+/* vec_sro */
+
+static vector signed char __ATTRS_o_ai
+vec_sro(vector signed char a, vector signed char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sro(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sro(vector unsigned char a, vector signed char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sro(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sro(vector short a, vector signed char b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sro(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sro(vector unsigned short a, vector signed char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sro(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sro(vector pixel a, vector signed char b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sro(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sro(vector int a, vector signed char b)
+{
+  return (vector int)__builtin_altivec_vsro(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sro(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsro(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sro(vector unsigned int a, vector signed char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sro(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_sro(vector float a, vector signed char b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_sro(vector float a, vector unsigned char b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+/* vec_vsro */
+
+static vector signed char __ATTRS_o_ai
+vec_vsro(vector signed char a, vector signed char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsro(vector signed char a, vector unsigned char b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsro(vector unsigned char a, vector signed char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsro(vector unsigned char a, vector unsigned char b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsro(vector short a, vector signed char b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsro(vector short a, vector unsigned char b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsro(vector unsigned short a, vector signed char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsro(vector unsigned short a, vector unsigned char b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsro(vector pixel a, vector signed char b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsro(vector pixel a, vector unsigned char b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsro(vector int a, vector signed char b)
+{
+  return (vector int)__builtin_altivec_vsro(a, (vector int)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsro(vector int a, vector unsigned char b)
+{
+  return (vector int)__builtin_altivec_vsro(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsro(vector unsigned int a, vector signed char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsro(vector unsigned int a, vector unsigned char b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsro(vector float a, vector signed char b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsro(vector float a, vector unsigned char b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)a, (vector int)b);
+}
+
+/* vec_st */
+
+static void __ATTRS_o_ai
+vec_st(vector signed char a, int b, vector signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned char a, int b, vector unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char a, int b, vector bool char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector short a, int b, vector short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned short a, int b, vector unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short a, int b, vector bool short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel a, int b, vector pixel *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector int a, int b, vector int *c)
+{
+  __builtin_altivec_stvx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned int a, int b, vector unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int a, int b, vector bool int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector float a, int b, vector float *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+/* vec_stvx */
+
+static void __ATTRS_o_ai
+vec_stvx(vector signed char a, int b, vector signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned char a, int b, vector unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char a, int b, vector bool char *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector short a, int b, vector short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned short a, int b, vector unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short a, int b, vector bool short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel a, int b, vector pixel *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector int a, int b, vector int *c)
+{
+  __builtin_altivec_stvx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned int a, int b, vector unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int a, int b, vector bool int *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector float a, int b, vector float *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvx((vector int)a, b, c);
+}
+
+/* vec_ste */
+
+static void __ATTRS_o_ai
+vec_ste(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvehx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvewx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+/* vec_stvebx */
+
+static void __ATTRS_o_ai
+vec_stvebx(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvebx((vector char)a, b, c);
+}
+
+/* vec_stvehx */
+
+static void __ATTRS_o_ai
+vec_stvehx(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvehx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvehx((vector short)a, b, c);
+}
+
+/* vec_stvewx */
+
+static void __ATTRS_o_ai
+vec_stvewx(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvewx(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvewx((vector int)a, b, c);
+}
+
+/* vec_stl */
+
+static void __ATTRS_o_ai
+vec_stl(vector signed char a, int b, vector signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned char a, int b, vector unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char a, int b, vector bool char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector short a, int b, vector short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned short a, int b, vector unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short a, int b, vector bool short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel a, int b, vector pixel *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector int a, int b, vector int *c)
+{
+  __builtin_altivec_stvxl(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvxl(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned int a, int b, vector unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int a, int b, vector bool int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector float a, int b, vector float *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+/* vec_stvxl */
+
+static void __ATTRS_o_ai
+vec_stvxl(vector signed char a, int b, vector signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector signed char a, int b, signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned char a, int b, vector unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char a, int b, signed char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char a, int b, unsigned char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char a, int b, vector bool char *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector short a, int b, vector short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector short a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned short a, int b, vector unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short a, int b, vector bool short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel a, int b, short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel a, int b, unsigned short *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel a, int b, vector pixel *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector int a, int b, vector int *c)
+{
+  __builtin_altivec_stvxl(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector int a, int b, int *c)
+{
+  __builtin_altivec_stvxl(a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned int a, int b, vector unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int a, int b, int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int a, int b, unsigned int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int a, int b, vector bool int *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector float a, int b, vector float *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector float a, int b, float *c)
+{
+  __builtin_altivec_stvxl((vector int)a, b, c);
+}
+
+/* vec_sub */
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector signed char a, vector signed char b)
+{
+  return a - b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a - b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector signed char a, vector bool char b)
+{
+  return a - (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector unsigned char a, vector unsigned char b)
+{
+  return a - b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a - b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector unsigned char a, vector bool char b)
+{
+  return a - (vector unsigned char)b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector short a, vector short b)
+{
+  return a - b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector bool short a, vector short b)
+{
+  return (vector short)a - b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector short a, vector bool short b)
+{
+  return a - (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector unsigned short a, vector unsigned short b)
+{
+  return a - b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a - b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector unsigned short a, vector bool short b)
+{
+  return a - (vector unsigned short)b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector int a, vector int b)
+{
+  return a - b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector bool int a, vector int b)
+{
+  return (vector int)a - b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector int a, vector bool int b)
+{
+  return a - (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector unsigned int a, vector unsigned int b)
+{
+  return a - b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a - b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector unsigned int a, vector bool int b)
+{
+  return a - (vector unsigned int)b;
+}
+
+static vector float __ATTRS_o_ai
+vec_sub(vector float a, vector float b)
+{
+  return a - b;
+}
+
+/* vec_vsububm */
+
+#define __builtin_altivec_vsububm vec_vsububm
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector signed char a, vector signed char b)
+{
+  return a - b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a - b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector signed char a, vector bool char b)
+{
+  return a - (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector unsigned char a, vector unsigned char b)
+{
+  return a - b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a - b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector unsigned char a, vector bool char b)
+{
+  return a - (vector unsigned char)b;
+}
+
+/* vec_vsubuhm */
+
+#define __builtin_altivec_vsubuhm vec_vsubuhm
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector short a, vector short b)
+{
+  return a - b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector bool short a, vector short b)
+{
+  return (vector short)a - b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector short a, vector bool short b)
+{
+  return a - (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector unsigned short a, vector unsigned short b)
+{
+  return a - b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a - b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector unsigned short a, vector bool short b)
+{
+  return a - (vector unsigned short)b;
+}
+
+/* vec_vsubuwm */
+
+#define __builtin_altivec_vsubuwm vec_vsubuwm
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector int a, vector int b)
+{
+  return a - b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector bool int a, vector int b)
+{
+  return (vector int)a - b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector int a, vector bool int b)
+{
+  return a - (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector unsigned int a, vector unsigned int b)
+{
+  return a - b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a - b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector unsigned int a, vector bool int b)
+{
+  return a - (vector unsigned int)b;
+}
+
+/* vec_vsubfp */
+
+#define __builtin_altivec_vsubfp vec_vsubfp
+
+static vector float __attribute__((__always_inline__))
+vec_vsubfp(vector float a, vector float b)
+{
+  return a - b;
+}
+
+/* vec_subc */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_subc(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubcuw(a, b);
+}
+
+/* vec_vsubcuw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vsubcuw(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubcuw(a, b);
+}
+
+/* vec_subs */
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vsubsbs(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vsubsbs((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vsubsbs(a, (vector signed char)b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vsububs(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vsububs((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vsububs(a, (vector unsigned char)b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector short a, vector short b)
+{
+  return __builtin_altivec_vsubshs(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vsubshs((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vsubshs(a, (vector short)b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsubuhs(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsubuhs((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vsubuhs(a, (vector unsigned short)b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector int a, vector int b)
+{
+  return __builtin_altivec_vsubsws(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vsubsws((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vsubsws(a, (vector int)b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubuws(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubuws((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vsubuws(a, (vector unsigned int)b);
+}
+
+/* vec_vsubsbs */
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vsubsbs(a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vsubsbs((vector signed char)a, b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vsubsbs(a, (vector signed char)b);
+}
+
+/* vec_vsububs */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vsububs(a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vsububs((vector unsigned char)a, b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vsububs(a, (vector unsigned char)b);
+}
+
+/* vec_vsubshs */
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector short a, vector short b)
+{
+  return __builtin_altivec_vsubshs(a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vsubshs((vector short)a, b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vsubshs(a, (vector short)b);
+}
+
+/* vec_vsubuhs */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsubuhs(a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vsubuhs((vector unsigned short)a, b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vsubuhs(a, (vector unsigned short)b);
+}
+
+/* vec_vsubsws */
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector int a, vector int b)
+{
+  return __builtin_altivec_vsubsws(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vsubsws((vector int)a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vsubsws(a, (vector int)b);
+}
+
+/* vec_vsubuws */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubuws(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vsubuws((vector unsigned int)a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vsubuws(a, (vector unsigned int)b);
+}
+
+/* vec_sum4s */
+
+static vector int __ATTRS_o_ai
+vec_sum4s(vector signed char a, vector int b)
+{
+  return __builtin_altivec_vsum4sbs(a, b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sum4s(vector unsigned char a, vector unsigned int b)
+{
+  return __builtin_altivec_vsum4ubs(a, b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sum4s(vector signed short a, vector int b)
+{
+  return __builtin_altivec_vsum4shs(a, b);
+}
+
+/* vec_vsum4sbs */
+
+static vector int __attribute__((__always_inline__))
+vec_vsum4sbs(vector signed char a, vector int b)
+{
+  return __builtin_altivec_vsum4sbs(a, b);
+}
+
+/* vec_vsum4ubs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vsum4ubs(vector unsigned char a, vector unsigned int b)
+{
+  return __builtin_altivec_vsum4ubs(a, b);
+}
+
+/* vec_vsum4shs */
+
+static vector int __attribute__((__always_inline__))
+vec_vsum4shs(vector signed short a, vector int b)
+{
+  return __builtin_altivec_vsum4shs(a, b);
+}
+
+/* vec_sum2s */
+
+static vector signed int __attribute__((__always_inline__))
+vec_sum2s(vector int a, vector int b)
+{
+  return __builtin_altivec_vsum2sws(a, b);
+}
+
+/* vec_vsum2sws */
+
+static vector signed int __attribute__((__always_inline__))
+vec_vsum2sws(vector int a, vector int b)
+{
+  return __builtin_altivec_vsum2sws(a, b);
+}
+
+/* vec_sums */
+
+static vector signed int __attribute__((__always_inline__))
+vec_sums(vector signed int a, vector signed int b)
+{
+  return __builtin_altivec_vsumsws(a, b);
+}
+
+/* vec_vsumsws */
+
+static vector signed int __attribute__((__always_inline__))
+vec_vsumsws(vector signed int a, vector signed int b)
+{
+  return __builtin_altivec_vsumsws(a, b);
+}
+
+/* vec_trunc */
+
+static vector float __attribute__((__always_inline__))
+vec_trunc(vector float a)
+{
+  return __builtin_altivec_vrfiz(a);
+}
+
+/* vec_vrfiz */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfiz(vector float a)
+{
+  return __builtin_altivec_vrfiz(a);
+}
+
+/* vec_unpackh */
+
+static vector short __ATTRS_o_ai
+vec_unpackh(vector signed char a)
+{
+  return __builtin_altivec_vupkhsb((vector char)a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_unpackh(vector bool char a)
+{
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)a);
+}
+
+static vector int __ATTRS_o_ai
+vec_unpackh(vector short a)
+{
+  return __builtin_altivec_vupkhsh(a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_unpackh(vector bool short a)
+{
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_unpackh(vector pixel a)
+{
+  return (vector unsigned int)__builtin_altivec_vupkhsh((vector short)a);
+}
+
+/* vec_vupkhsb */
+
+static vector short __ATTRS_o_ai
+vec_vupkhsb(vector signed char a)
+{
+  return __builtin_altivec_vupkhsb((vector char)a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vupkhsb(vector bool char a)
+{
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)a);
+}
+
+/* vec_vupkhsh */
+
+static vector int __ATTRS_o_ai
+vec_vupkhsh(vector short a)
+{
+  return __builtin_altivec_vupkhsh(a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vupkhsh(vector bool short a)
+{
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vupkhsh(vector pixel a)
+{
+  return (vector unsigned int)__builtin_altivec_vupkhsh((vector short)a);
+}
+
+/* vec_unpackl */
+
+static vector short __ATTRS_o_ai
+vec_unpackl(vector signed char a)
+{
+  return __builtin_altivec_vupklsb((vector char)a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_unpackl(vector bool char a)
+{
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)a);
+}
+
+static vector int __ATTRS_o_ai
+vec_unpackl(vector short a)
+{
+  return __builtin_altivec_vupklsh(a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_unpackl(vector bool short a)
+{
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_unpackl(vector pixel a)
+{
+  return (vector unsigned int)__builtin_altivec_vupklsh((vector short)a);
+}
+
+/* vec_vupklsb */
+
+static vector short __ATTRS_o_ai
+vec_vupklsb(vector signed char a)
+{
+  return __builtin_altivec_vupklsb((vector char)a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vupklsb(vector bool char a)
+{
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)a);
+}
+
+/* vec_vupklsh */
+
+static vector int __ATTRS_o_ai
+vec_vupklsh(vector short a)
+{
+  return __builtin_altivec_vupklsh(a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vupklsh(vector bool short a)
+{
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vupklsh(vector pixel a)
+{
+  return (vector unsigned int)__builtin_altivec_vupklsh((vector short)a);
+}
+
+/* vec_xor */
+
+#define __builtin_altivec_vxor vec_xor
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector signed char a, vector signed char b)
+{
+  return a ^ b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a ^ b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector signed char a, vector bool char b)
+{
+  return a ^ (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector unsigned char a, vector unsigned char b)
+{
+  return a ^ b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a ^ b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector unsigned char a, vector bool char b)
+{
+  return a ^ (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_xor(vector bool char a, vector bool char b)
+{
+  return a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector short a, vector short b)
+{
+  return a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector bool short a, vector short b)
+{
+  return (vector short)a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector short a, vector bool short b)
+{
+  return a ^ (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector unsigned short a, vector unsigned short b)
+{
+  return a ^ b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a ^ b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector unsigned short a, vector bool short b)
+{
+  return a ^ (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_xor(vector bool short a, vector bool short b)
+{
+  return a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector int a, vector int b)
+{
+  return a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector bool int a, vector int b)
+{
+  return (vector int)a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector int a, vector bool int b)
+{
+  return a ^ (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector unsigned int a, vector unsigned int b)
+{
+  return a ^ b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a ^ b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector unsigned int a, vector bool int b)
+{
+  return a ^ (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_xor(vector bool int a, vector bool int b)
+{
+  return a ^ b;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* vec_vxor */
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector signed char a, vector signed char b)
+{
+  return a ^ b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector bool char a, vector signed char b)
+{
+  return (vector signed char)a ^ b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector signed char a, vector bool char b)
+{
+  return a ^ (vector signed char)b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector unsigned char a, vector unsigned char b)
+{
+  return a ^ b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector bool char a, vector unsigned char b)
+{
+  return (vector unsigned char)a ^ b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector unsigned char a, vector bool char b)
+{
+  return a ^ (vector unsigned char)b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vxor(vector bool char a, vector bool char b)
+{
+  return a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector short a, vector short b)
+{
+  return a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector bool short a, vector short b)
+{
+  return (vector short)a ^ b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector short a, vector bool short b)
+{
+  return a ^ (vector short)b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector unsigned short a, vector unsigned short b)
+{
+  return a ^ b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector bool short a, vector unsigned short b)
+{
+  return (vector unsigned short)a ^ b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector unsigned short a, vector bool short b)
+{
+  return a ^ (vector unsigned short)b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vxor(vector bool short a, vector bool short b)
+{
+  return a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector int a, vector int b)
+{
+  return a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector bool int a, vector int b)
+{
+  return (vector int)a ^ b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector int a, vector bool int b)
+{
+  return a ^ (vector int)b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector unsigned int a, vector unsigned int b)
+{
+  return a ^ b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector bool int a, vector unsigned int b)
+{
+  return (vector unsigned int)a ^ b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector unsigned int a, vector bool int b)
+{
+  return a ^ (vector unsigned int)b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vxor(vector bool int a, vector bool int b)
+{
+  return a ^ b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector float a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector bool int a, vector float b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector float a, vector bool int b)
+{
+  vector unsigned int res = (vector unsigned int)a ^ (vector unsigned int)b;
+  return (vector float)res;
+}
+
+/* ------------------------ extensions for CBEA ----------------------------- */
+
+/* vec_extract */
+
+static signed char __ATTRS_o_ai
+vec_extract(vector signed char a, int b)
+{
+  return a[b];
+}
+
+static unsigned char __ATTRS_o_ai
+vec_extract(vector unsigned char a, int b)
+{
+  return a[b];
+}
+
+static short __ATTRS_o_ai
+vec_extract(vector short a, int b)
+{
+  return a[b];
+}
+
+static unsigned short __ATTRS_o_ai
+vec_extract(vector unsigned short a, int b)
+{
+  return a[b];
+}
+
+static int __ATTRS_o_ai
+vec_extract(vector int a, int b)
+{
+  return a[b];
+}
+
+static unsigned int __ATTRS_o_ai
+vec_extract(vector unsigned int a, int b)
+{
+  return a[b];
+}
+
+static float __ATTRS_o_ai
+vec_extract(vector float a, int b)
+{
+  return a[b];
+}
+
+/* vec_insert */
+
+static vector signed char __ATTRS_o_ai
+vec_insert(signed char a, vector signed char b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_insert(unsigned char a, vector unsigned char b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector short __ATTRS_o_ai
+vec_insert(short a, vector short b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_insert(unsigned short a, vector unsigned short b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector int __ATTRS_o_ai
+vec_insert(int a, vector int b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_insert(unsigned int a, vector unsigned int b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+static vector float __ATTRS_o_ai
+vec_insert(float a, vector float b, int c)
+{
+  b[c] = a;
+  return b;
+}
+
+/* vec_lvlx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvlx(int a, const signed char *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector signed char)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvlx(int a, const vector signed char *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector signed char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlx(int a, const unsigned char *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlx(int a, const vector unsigned char *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector unsigned char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvlx(int a, const vector bool char *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector bool char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlx(int a, const short *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector short)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlx(int a, const vector short *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlx(int a, const unsigned short *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlx(int a, const vector unsigned short *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector unsigned short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvlx(int a, const vector bool short *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector bool short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvlx(int a, const vector pixel *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector pixel)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlx(int a, const int *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector int)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlx(int a, const vector int *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlx(int a, const unsigned int *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlx(int a, const vector unsigned int *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector unsigned int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvlx(int a, const vector bool int *b)
+{
+  return vec_perm(vec_ld(a, b), 
+                  (vector bool int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlx(int a, const float *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector float)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlx(int a, const vector float *b)
+{
+  return vec_perm(vec_ld(a, b),
+                  (vector float)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+/* vec_lvlxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvlxl(int a, const signed char *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector signed char)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvlxl(int a, const vector signed char *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector signed char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlxl(int a, const unsigned char *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlxl(int a, const vector unsigned char *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector unsigned char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvlxl(int a, const vector bool char *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector bool char)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlxl(int a, const short *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector short)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlxl(int a, const vector short *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlxl(int a, const unsigned short *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlxl(int a, const vector unsigned short *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector unsigned short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvlxl(int a, const vector bool short *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector bool short)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvlxl(int a, const vector pixel *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector pixel)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlxl(int a, const int *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector int)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlxl(int a, const vector int *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlxl(int a, const unsigned int *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlxl(int a, const vector unsigned int *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector unsigned int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvlxl(int a, const vector bool int *b)
+{
+  return vec_perm(vec_ldl(a, b), 
+                  (vector bool int)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlxl(int a, const float *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector float)(0),
+                  vec_lvsl(a, b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlxl(int a, vector float *b)
+{
+  return vec_perm(vec_ldl(a, b),
+                  (vector float)(0),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+/* vec_lvrx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvrx(int a, const signed char *b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvrx(int a, const vector signed char *b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrx(int a, const unsigned char *b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrx(int a, const vector unsigned char *b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvrx(int a, const vector bool char *b)
+{
+  return vec_perm((vector bool char)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrx(int a, const short *b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrx(int a, const vector short *b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrx(int a, const unsigned short *b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrx(int a, const vector unsigned short *b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvrx(int a, const vector bool short *b)
+{
+  return vec_perm((vector bool short)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvrx(int a, const vector pixel *b)
+{
+  return vec_perm((vector pixel)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrx(int a, const int *b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrx(int a, const vector int *b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrx(int a, const unsigned int *b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrx(int a, const vector unsigned int *b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvrx(int a, const vector bool int *b)
+{
+  return vec_perm((vector bool int)(0),
+                  vec_ld(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrx(int a, const float *b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrx(int a, const vector float *b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ld(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+/* vec_lvrxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvrxl(int a, const signed char *b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvrxl(int a, const vector signed char *b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrxl(int a, const unsigned char *b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrxl(int a, const vector unsigned char *b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvrxl(int a, const vector bool char *b)
+{
+  return vec_perm((vector bool char)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrxl(int a, const short *b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrxl(int a, const vector short *b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrxl(int a, const unsigned short *b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrxl(int a, const vector unsigned short *b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvrxl(int a, const vector bool short *b)
+{
+  return vec_perm((vector bool short)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvrxl(int a, const vector pixel *b)
+{
+  return vec_perm((vector pixel)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrxl(int a, const int *b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrxl(int a, const vector int *b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrxl(int a, const unsigned int *b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrxl(int a, const vector unsigned int *b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvrxl(int a, const vector bool int *b)
+{
+  return vec_perm((vector bool int)(0),
+                  vec_ldl(a, b), 
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrxl(int a, const float *b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrxl(int a, const vector float *b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ldl(a, b),
+                  vec_lvsl(a, (unsigned char *)b));
+}
+
+/* vec_stvlx */
+
+static void __ATTRS_o_ai
+vec_stvlx(vector signed char a, int b, signed char *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector signed char a, int b, vector signed char *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned char a, int b, unsigned char *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned char a, int b, vector unsigned char *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool char a, int b, vector bool char *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector short a, int b, short *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector short a, int b, vector short *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned short a, int b, unsigned short *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned short a, int b, vector unsigned short *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool short a, int b, vector bool short *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector pixel a, int b, vector pixel *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector int a, int b, int *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector int a, int b, vector int *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned int a, int b, unsigned int *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned int a, int b, vector unsigned int *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool int a, int b, vector bool int *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector float a, int b, vector float *c)
+{
+  return vec_st(vec_perm(vec_lvrx(b, c),
+                         a,
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+/* vec_stvlxl */
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector signed char a, int b, signed char *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector signed char a, int b, vector signed char *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned char a, int b, unsigned char *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned char a, int b, vector unsigned char *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool char a, int b, vector bool char *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector short a, int b, short *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector short a, int b, vector short *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned short a, int b, unsigned short *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned short a, int b, vector unsigned short *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool short a, int b, vector bool short *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector pixel a, int b, vector pixel *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector int a, int b, int *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector int a, int b, vector int *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned int a, int b, unsigned int *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned int a, int b, vector unsigned int *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool int a, int b, vector bool int *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector float a, int b, vector float *c)
+{
+  return vec_stl(vec_perm(vec_lvrx(b, c),
+                          a,
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+/* vec_stvrx */
+
+static void __ATTRS_o_ai
+vec_stvrx(vector signed char a, int b, signed char *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector signed char a, int b, vector signed char *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned char a, int b, unsigned char *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned char a, int b, vector unsigned char *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool char a, int b, vector bool char *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector short a, int b, short *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector short a, int b, vector short *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned short a, int b, unsigned short *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned short a, int b, vector unsigned short *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool short a, int b, vector bool short *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector pixel a, int b, vector pixel *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector int a, int b, int *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector int a, int b, vector int *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned int a, int b, unsigned int *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned int a, int b, vector unsigned int *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool int a, int b, vector bool int *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector float a, int b, vector float *c)
+{
+  return vec_st(vec_perm(a,
+                         vec_lvlx(b, c),
+                         vec_lvsr(b, (unsigned char *)c)),
+                b, c);
+}
+
+/* vec_stvrxl */
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector signed char a, int b, signed char *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector signed char a, int b, vector signed char *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned char a, int b, unsigned char *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned char a, int b, vector unsigned char *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool char a, int b, vector bool char *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector short a, int b, short *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector short a, int b, vector short *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned short a, int b, unsigned short *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned short a, int b, vector unsigned short *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool short a, int b, vector bool short *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector pixel a, int b, vector pixel *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector int a, int b, int *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector int a, int b, vector int *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned int a, int b, unsigned int *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned int a, int b, vector unsigned int *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool int a, int b, vector bool int *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector float a, int b, vector float *c)
+{
+  return vec_stl(vec_perm(a,
+                          vec_lvlx(b, c),
+                          vec_lvsr(b, (unsigned char *)c)),
+                 b, c);
+}
+
+/* vec_promote */
+
+static vector signed char __ATTRS_o_ai
+vec_promote(signed char a, int b)
+{
+  vector signed char res = (vector signed char)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_promote(unsigned char a, int b)
+{
+  vector unsigned char res = (vector unsigned char)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector short __ATTRS_o_ai
+vec_promote(short a, int b)
+{
+  vector short res = (vector short)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_promote(unsigned short a, int b)
+{
+  vector unsigned short res = (vector unsigned short)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector int __ATTRS_o_ai
+vec_promote(int a, int b)
+{
+  vector int res = (vector int)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_promote(unsigned int a, int b)
+{
+  vector unsigned int res = (vector unsigned int)(0);
+  res[b] = a;
+  return res;
+}
+
+static vector float __ATTRS_o_ai
+vec_promote(float a, int b)
+{
+  vector float res = (vector float)(0);
+  res[b] = a;
+  return res;
+}
+
+/* vec_splats */
+
+static vector signed char __ATTRS_o_ai
+vec_splats(signed char a)
+{
+  return (vector signed char)(a);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_splats(unsigned char a)
+{
+  return (vector unsigned char)(a);
+}
+
+static vector short __ATTRS_o_ai
+vec_splats(short a)
+{
+  return (vector short)(a);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_splats(unsigned short a)
+{
+  return (vector unsigned short)(a);
+}
+
+static vector int __ATTRS_o_ai
+vec_splats(int a)
+{
+  return (vector int)(a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_splats(unsigned int a)
+{
+  return (vector unsigned int)(a);
+}
+
+static vector float __ATTRS_o_ai
+vec_splats(float a)
+{
+  return (vector float)(a);
+}
+
+/* ----------------------------- predicates --------------------------------- */
+
+/* vec_all_eq */
+
+static int __ATTRS_o_ai
+vec_all_eq(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned short a, vector unsigned short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned short a, vector bool short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short a, vector short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short a, vector unsigned short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short a, vector bool short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector pixel a, vector pixel b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT, a, b);
+}
+
+/* vec_all_ge */
+
+static int __ATTRS_o_ai
+vec_all_ge(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, (vector signed char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, (vector unsigned char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, b, (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, (vector short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, (vector unsigned short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, b, (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, (vector int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, (vector unsigned int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, b, (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT, a, b);
+}
+
+/* vec_all_gt */
+
+static int __ATTRS_o_ai
+vec_all_gt(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, a, (vector signed char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, a, (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, (vector unsigned char)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, a, (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, (vector unsigned short)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, a, (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, (vector unsigned int)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT, a, b);
+}
+
+/* vec_all_in */
+
+static int __attribute__((__always_inline__))
+vec_all_in(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpbfp_p(__CR6_EQ, a, b);
+}
+
+/* vec_all_le */
+
+static int __ATTRS_o_ai
+vec_all_le(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, a, (vector signed char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, a, (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, (vector unsigned char)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, a, (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, (vector unsigned short)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, a, (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, (vector unsigned int)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT, b, a);
+}
+
+/* vec_all_lt */
+
+static int __ATTRS_o_ai
+vec_all_lt(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, (vector signed char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, (vector unsigned char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, b, (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, (vector short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, (vector unsigned short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, b, (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, (vector int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, (vector unsigned int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, b, (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT, b, a);
+}
+
+/* vec_all_nan */
+
+static int __attribute__((__always_inline__))
+vec_all_nan(vector float a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ, a, a);
+}
+
+/* vec_all_ne */
+
+static int __ATTRS_o_ai
+vec_all_ne(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned short a, vector unsigned short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned short a, vector bool short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short a, vector short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short a, vector unsigned short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short a, vector bool short b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector pixel a, vector pixel b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ, a, b);
+}
+
+/* vec_all_nge */
+
+static int __attribute__((__always_inline__))
+vec_all_nge(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ, a, b);
+}
+
+/* vec_all_ngt */
+
+static int __attribute__((__always_inline__))
+vec_all_ngt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ, a, b);
+}
+
+/* vec_all_nle */
+
+static int __attribute__((__always_inline__))
+vec_all_nle(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ, b, a);
+}
+
+/* vec_all_nlt */
+
+static int __attribute__((__always_inline__))
+vec_all_nlt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ, b, a);
+}
+
+/* vec_all_numeric */
+
+static int __attribute__((__always_inline__))
+vec_all_numeric(vector float a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT, a, a);
+}
+
+/* vec_any_eq */
+
+static int __ATTRS_o_ai
+vec_any_eq(vector signed char a, vector signed char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector signed char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned char a, vector unsigned char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char a, vector signed char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char a, vector unsigned char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector pixel a, vector pixel b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned int a, vector unsigned int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned int a, vector bool int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int a, vector int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int a, vector unsigned int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int a, vector bool int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ_REV, a, b);
+}
+
+/* vec_any_ge */
+
+static int __ATTRS_o_ai
+vec_any_ge(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, (vector signed char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, (vector unsigned char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, b, (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, (vector short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned short a, vector bool short b)
+{
+  return
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, (vector unsigned short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short a, vector unsigned short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, b, (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, (vector int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, (vector unsigned int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, b, (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ_REV, a, b);
+}
+
+/* vec_any_gt */
+
+static int __ATTRS_o_ai
+vec_any_gt(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, a, (vector signed char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned char a, vector bool char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, a, (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char a, vector unsigned char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, (vector unsigned char)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned short a, vector bool short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, a, (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short a, vector unsigned short b)
+{
+  return
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, (vector unsigned short)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, a, (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, (vector unsigned int)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ_REV, a, b);
+}
+
+/* vec_any_le */
+
+static int __ATTRS_o_ai
+vec_any_le(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, a, (vector signed char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned char a, vector bool char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, a, (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char a, vector unsigned char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, (vector unsigned char)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)a,
+                                      (vector unsigned char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned short a, vector bool short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, a, (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short a, vector unsigned short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, (vector unsigned short)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)a,
+                                      (vector unsigned short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, a, (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, (vector unsigned int)a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)a,
+                                      (vector unsigned int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ_REV, b, a);
+}
+
+/* vec_any_lt */
+
+static int __ATTRS_o_ai
+vec_any_lt(vector signed char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector signed char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, (vector signed char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned char a, vector unsigned char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned char a, vector bool char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, (vector unsigned char)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char a, vector signed char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char a, vector unsigned char b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, b, (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char a, vector bool char b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)b,
+                                      (vector unsigned char)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, (vector short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned short a, vector bool short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, (vector unsigned short)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short a, vector unsigned short b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, b, (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)b,
+                                      (vector unsigned short)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, (vector int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, (vector unsigned int)b, a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int a, vector int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int a, vector unsigned int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, b, (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)b,
+                                      (vector unsigned int)a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ_REV, b, a);
+}
+
+/* vec_any_nan */
+
+static int __attribute__((__always_inline__))
+vec_any_nan(vector float a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT_REV, a, a);
+}
+
+/* vec_any_ne */
+
+static int __ATTRS_o_ai
+vec_any_ne(vector signed char a, vector signed char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector signed char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned char a, vector unsigned char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char a, vector signed char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char a, vector unsigned char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char a, vector bool char b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)a, (vector char)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, a, (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, 
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short a, vector short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short a, vector unsigned short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short a, vector bool short b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector pixel a, vector pixel b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)a,
+                                      (vector short)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector int a, vector int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT_REV, a, b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector int a, vector bool int b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT_REV, a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned int a, vector unsigned int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned int a, vector bool int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int a, vector int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int a, vector unsigned int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int a, vector bool int b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)a, (vector int)b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT_REV, a, b);
+}
+
+/* vec_any_nge */
+
+static int __attribute__((__always_inline__))
+vec_any_nge(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT_REV, a, b);
+}
+
+/* vec_any_ngt */
+
+static int __attribute__((__always_inline__))
+vec_any_ngt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT_REV, a, b);
+}
+
+/* vec_any_nle */
+
+static int __attribute__((__always_inline__))
+vec_any_nle(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT_REV, b, a);
+}
+
+/* vec_any_nlt */
+
+static int __attribute__((__always_inline__))
+vec_any_nlt(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT_REV, b, a);
+}
+
+/* vec_any_numeric */
+
+static int __attribute__((__always_inline__))
+vec_any_numeric(vector float a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ_REV, a, a);
+}
+
+/* vec_any_out */
+
+static int __attribute__((__always_inline__))
+vec_any_out(vector float a, vector float b)
+{
+  return __builtin_altivec_vcmpbfp_p(__CR6_EQ_REV, a, b);
+}
+
+#undef __ATTRS_o_ai
+
+#endif /* __ALTIVEC_H */
diff --git a/clang/lib/Headers/avx2intrin.h b/clang/lib/Headers/avx2intrin.h
new file mode 100644
index 0000000..884c46d
--- /dev/null
+++ b/clang/lib/Headers/avx2intrin.h
@@ -0,0 +1,961 @@
+/*===---- avx2intrin.h - AVX2 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avx2intrin.h> directly; include <immintrin.h> instead."
+#endif
+
+/* SSE4 Multiple Packed Sums of Absolute Difference.  */
+#define _mm256_mpsadbw_epu8(X, Y, M) __builtin_ia32_mpsadbw256((X), (Y), (M))
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi8(__m256i a)
+{
+    return (__m256i)__builtin_ia32_pabsb256((__v32qi)a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi16(__m256i a)
+{
+    return (__m256i)__builtin_ia32_pabsw256((__v16hi)a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi32(__m256i a)
+{
+    return (__m256i)__builtin_ia32_pabsd256((__v8si)a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packs_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_packsswb256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packs_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_packssdw256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packus_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_packuswb256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packus_epi32(__m256i __V1, __m256i __V2)
+{
+  return (__m256i) __builtin_ia32_packusdw256((__v8si)__V1, (__v8si)__V2);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)((__v32qi)a + (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)((__v16hi)a + (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)((__v8si)a + (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi64(__m256i a, __m256i b)
+{
+  return a + b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_paddsb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_paddsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epu8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_paddusb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_paddusw256((__v16hi)a, (__v16hi)b);
+}
+
+#define _mm256_alignr_epi8(a, b, n) __extension__ ({ \
+  __m256i __a = (a); \
+  __m256i __b = (b); \
+  (__m256i)__builtin_ia32_palignr256((__v32qi)__a, (__v32qi)__b, (n)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_and_si256(__m256i a, __m256i b)
+{
+  return a & b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_si256(__m256i a, __m256i b)
+{
+  return ~a & b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_avg_epu8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pavgb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_avg_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pavgw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_epi8(__m256i __V1, __m256i __V2, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_pblendvb256((__v32qi)__V1, (__v32qi)__V2,
+                                              (__v32qi)__M);
+}
+
+#define _mm256_blend_epi16(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_pblendw256((__v16hi)__V1, (__v16hi)__V2, (M)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)((__v32qi)a == (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)((__v16hi)a == (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)((__v8si)a == (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi64(__m256i a, __m256i b)
+{
+  return (__m256i)(a == b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)((__v32qi)a > (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)((__v16hi)a > (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)((__v8si)a > (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi64(__m256i a, __m256i b)
+{
+  return (__m256i)(a > b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phaddw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_epi32(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phaddd256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadds_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phaddsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phsubw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_epi32(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phsubd256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsubs_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_phsubsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maddubs_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_pmaddubsw256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_madd_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaddwd256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxsb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxsd256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxub256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxuw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmaxud256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminsb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminsd256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminub256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminuw256 ((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pminud256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_epi8(__m256i a)
+{
+  return __builtin_ia32_pmovmskb256((__v32qi)a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi16(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbw256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbd256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbq256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi16_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxwd256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi16_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxwq256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxdq256((__v4si)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi16(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbw256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbd256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbq256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu16_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxwd256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu16_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxwq256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu32_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxdq256((__v4si)__V);
+}
+
+static __inline__  __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmuldq256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhrs_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmulhrsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhi_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmulhuw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhi_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pmulhw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mullo_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)((__v16hi)a * (__v16hi)b);
+}
+
+static __inline__  __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mullo_epi32 (__m256i a, __m256i b)
+{
+  return (__m256i)((__v8si)a * (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_epu32(__m256i a, __m256i b)
+{
+  return __builtin_ia32_pmuludq256((__v8si)a, (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_or_si256(__m256i a, __m256i b)
+{
+  return a | b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sad_epu8(__m256i a, __m256i b)
+{
+  return __builtin_ia32_psadbw256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_shuffle_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_pshufb256((__v32qi)a, (__v32qi)b);
+}
+
+#define _mm256_shuffle_epi32(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v8si)__a, (__v8si)_mm256_set1_epi32(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6)); })
+
+#define _mm256_shufflehi_epi16(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)_mm256_set1_epi16(0), \
+                                   0, 1, 2, 3, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6), \
+                                   8, 9, 10, 11, \
+                                   12 + (((imm) & 0x03) >> 0), \
+                                   12 + (((imm) & 0x0c) >> 2), \
+                                   12 + (((imm) & 0x30) >> 4), \
+                                   12 + (((imm) & 0xc0) >> 6)); })
+
+#define _mm256_shufflelo_epi16(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)_mm256_set1_epi16(0), \
+                                   (imm) & 0x3,((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4, 5, 6, 7, \
+                                   8 + (((imm) & 0x03) >> 0), \
+                                   8 + (((imm) & 0x0c) >> 2), \
+                                   8 + (((imm) & 0x30) >> 4), \
+                                   8 + (((imm) & 0xc0) >> 6), \
+                                   12, 13, 14, 15); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi8(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_psignb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi16(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_psignw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi32(__m256i a, __m256i b)
+{
+    return (__m256i)__builtin_ia32_psignd256((__v8si)a, (__v8si)b);
+}
+
+#define _mm256_slli_si256(a, count) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_ia32_pslldqi256(__a, (count)*8); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi16(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_psllwi256((__v16hi)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi16(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_psllw256((__v16hi)a, (__v8hi)count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi32(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_pslldi256((__v8si)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi32(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_pslld256((__v8si)a, (__v4si)count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi64(__m256i a, int count)
+{
+  return __builtin_ia32_psllqi256(a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi64(__m256i a, __m128i count)
+{
+  return __builtin_ia32_psllq256(a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srai_epi16(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_psrawi256((__v16hi)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sra_epi16(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_psraw256((__v16hi)a, (__v8hi)count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srai_epi32(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_psradi256((__v8si)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sra_epi32(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_psrad256((__v8si)a, (__v4si)count);
+}
+
+#define _mm256_srli_si256(a, count) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_ia32_psrldqi256(__a, (count)*8); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi16(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_psrlwi256((__v16hi)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi16(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_psrlw256((__v16hi)a, (__v8hi)count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi32(__m256i a, int count)
+{
+  return (__m256i)__builtin_ia32_psrldi256((__v8si)a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi32(__m256i a, __m128i count)
+{
+  return (__m256i)__builtin_ia32_psrld256((__v8si)a, (__v4si)count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi64(__m256i a, int count)
+{
+  return __builtin_ia32_psrlqi256(a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi64(__m256i a, __m128i count)
+{
+  return __builtin_ia32_psrlq256(a, count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)((__v32qi)a - (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)((__v16hi)a - (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)((__v8si)a - (__v8si)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi64(__m256i a, __m256i b)
+{
+  return a - b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_psubsb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_psubsw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epu8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_psubusb256((__v32qi)a, (__v32qi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epu16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_psubusw256((__v16hi)a, (__v16hi)b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v32qi)a, (__v32qi)b, 8, 32+8, 9, 32+9, 10, 32+10, 11, 32+11, 12, 32+12, 13, 32+13, 14, 32+14, 15, 32+15, 24, 32+24, 25, 32+25, 26, 32+26, 27, 32+27, 28, 32+28, 29, 32+29, 30, 32+30, 31, 32+31);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v16hi)a, (__v16hi)b, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v8si)a, (__v8si)b, 2, 8+2, 3, 8+3, 6, 8+6, 7, 8+7);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi64(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector(a, b, 1, 4+1, 3, 4+3);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi8(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v32qi)a, (__v32qi)b, 0, 32+0, 1, 32+1, 2, 32+2, 3, 32+3, 4, 32+4, 5, 32+5, 6, 32+6, 7, 32+7, 16, 32+16, 17, 32+17, 18, 32+18, 19, 32+19, 20, 32+20, 21, 32+21, 22, 32+22, 23, 32+23);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi16(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v16hi)a, (__v16hi)b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector((__v8si)a, (__v8si)b, 0, 8+0, 1, 8+1, 4, 8+4, 5, 8+5);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi64(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_shufflevector(a, b, 0, 4+0, 2, 4+2);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_si256(__m256i a, __m256i b)
+{
+  return a ^ b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_load_si256(__m256i *__V)
+{
+  return (__m256i)__builtin_ia32_movntdqa256((__v4di *)__V);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastss_ps(__m128 __X)
+{
+  return (__m128)__builtin_ia32_vbroadcastss_ps((__v4sf)__X);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastss_ps(__m128 __X)
+{
+  return (__m256)__builtin_ia32_vbroadcastss_ps256((__v4sf)__X);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastsd_pd(__m128d __X)
+{
+  return (__m256d)__builtin_ia32_vbroadcastsd_pd256((__v2df)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastsi128_si256(__m128i const *a)
+{
+  return (__m256i)__builtin_ia32_vbroadcastsi256(a);
+}
+
+#define _mm_blend_epi32(V1, V2, M) __extension__ ({ \
+  __m128i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m128i)__builtin_ia32_pblendd128((__v4si)__V1, (__v4si)__V2, (M)); })
+
+#define _mm256_blend_epi32(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_pblendd256((__v8si)__V1, (__v8si)__V2, (M)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastb_epi8(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastb256((__v16qi)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastw_epi16(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastw256((__v8hi)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastd_epi32(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastd256((__v4si)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastq_epi64(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastq256(__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastb_epi8(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastb128((__v16qi)__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastw_epi16(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastw128((__v8hi)__X);
+}
+
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastd_epi32(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastd128((__v4si)__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastq_epi64(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastq128(__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar8x32_epi32(__m256i a, __m256i b)
+{
+  return (__m256i)__builtin_ia32_permvarsi256((__v8si)a, (__v8si)b);
+}
+
+#define _mm256_permute4x64_pd(V, M) __extension__ ({ \
+  __m256d __V = (V); \
+  (__m256d)__builtin_shufflevector((__v4df)__V, (__v4df) _mm256_setzero_pd(), \
+                                   (M) & 0x3, ((M) & 0xc) >> 2, \
+                                   ((M) & 0x30) >> 4, ((M) & 0xc0) >> 6); })
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar8x32_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_permvarsf256((__v8sf)a, (__v8sf)b);
+}
+
+#define _mm256_permute4x64_epi64(V, M) __extension__ ({ \
+  __m256i __V = (V); \
+  (__m256i)__builtin_shufflevector((__v4di)__V, (__v4di) _mm256_setzero_si256(), \
+                                   (M) & 0x3, ((M) & 0xc) >> 2, \
+                                   ((M) & 0x30) >> 4, ((M) & 0xc0) >> 6); })
+
+#define _mm256_permute2x128_si256(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_permti256(__V1, __V2, (M)); })
+
+#define _mm256_extracti128_si256(A, O) __extension__ ({ \
+  __m256i __A = (A); \
+  (__m128i)__builtin_ia32_extract128i256(__A, (O)); })
+
+#define _mm256_inserti128_si256(V1, V2, O) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m256i)__builtin_ia32_insert128i256(__V1, __V2, (O)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_epi32(int const *__X, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_maskloadd256((const __v8si *)__X, (__v8si)__M);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_epi64(long long const *__X, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_maskloadq256((const __v4di *)__X, __M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_epi32(int const *__X, __m128i __M)
+{
+  return (__m128i)__builtin_ia32_maskloadd((const __v4si *)__X, (__v4si)__M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_epi64(long long const *__X, __m128i __M)
+{
+  return (__m128i)__builtin_ia32_maskloadq((const __v2di *)__X, (__v2di)__M);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_epi32(int *__X, __m256i __M, __m256i __Y)
+{
+  __builtin_ia32_maskstored256((__v8si *)__X, (__v8si)__M, (__v8si)__Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_epi64(long long *__X, __m256i __M, __m256i __Y)
+{
+  __builtin_ia32_maskstoreq256((__v4di *)__X, __M, __Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_epi32(int *__X, __m128i __M, __m128i __Y)
+{
+  __builtin_ia32_maskstored((__v4si *)__X, (__v4si)__M, (__v4si)__Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_epi64(long long *__X, __m128i __M, __m128i __Y)
+{
+  __builtin_ia32_maskstoreq(( __v2di *)__X, __M, __Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sllv_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psllv8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sllv_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psllv4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sllv_epi64(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psllv4di(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sllv_epi64(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psllv2di(__X, __Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srav_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrav8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srav_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrav4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srlv_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrlv8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srlv_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrlv4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srlv_epi64(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrlv4di(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srlv_epi64(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrlv2di(__X, __Y);
+}
diff --git a/clang/lib/Headers/avxintrin.h b/clang/lib/Headers/avxintrin.h
new file mode 100644
index 0000000..ee7f835
--- /dev/null
+++ b/clang/lib/Headers/avxintrin.h
@@ -0,0 +1,1215 @@
+/*===---- avxintrin.h - AVX intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avxintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+typedef double __v4df __attribute__ ((__vector_size__ (32)));
+typedef float __v8sf __attribute__ ((__vector_size__ (32)));
+typedef long long __v4di __attribute__ ((__vector_size__ (32)));
+typedef int __v8si __attribute__ ((__vector_size__ (32)));
+typedef short __v16hi __attribute__ ((__vector_size__ (32)));
+typedef char __v32qi __attribute__ ((__vector_size__ (32)));
+
+typedef float __m256 __attribute__ ((__vector_size__ (32)));
+typedef double __m256d __attribute__((__vector_size__(32)));
+typedef long long __m256i __attribute__((__vector_size__(32)));
+
+/* Arithmetic */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_add_pd(__m256d a, __m256d b)
+{
+  return a+b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_add_ps(__m256 a, __m256 b)
+{
+  return a+b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_pd(__m256d a, __m256d b)
+{
+  return a-b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_ps(__m256 a, __m256 b)
+{
+  return a-b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_addsub_pd(__m256d a, __m256d b)
+{
+  return (__m256d)__builtin_ia32_addsubpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_addsub_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_addsubps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_div_pd(__m256d a, __m256d b)
+{
+  return a / b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_div_ps(__m256 a, __m256 b)
+{
+  return a / b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_max_pd(__m256d a, __m256d b)
+{
+  return (__m256d)__builtin_ia32_maxpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_max_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_maxps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_min_pd(__m256d a, __m256d b)
+{
+  return (__m256d)__builtin_ia32_minpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_min_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_minps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_pd(__m256d a, __m256d b)
+{
+  return a * b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_ps(__m256 a, __m256 b)
+{
+  return a * b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_sqrt_pd(__m256d a)
+{
+  return (__m256d)__builtin_ia32_sqrtpd256((__v4df)a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_sqrt_ps(__m256 a)
+{
+  return (__m256)__builtin_ia32_sqrtps256((__v8sf)a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_rsqrt_ps(__m256 a)
+{
+  return (__m256)__builtin_ia32_rsqrtps256((__v8sf)a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_rcp_ps(__m256 a)
+{
+  return (__m256)__builtin_ia32_rcpps256((__v8sf)a);
+}
+
+#define _mm256_round_pd(V, M) __extension__ ({ \
+    __m256d __V = (V); \
+    (__m256d)__builtin_ia32_roundpd256((__v4df)__V, (M)); })
+
+#define _mm256_round_ps(V, M) __extension__ ({ \
+  __m256 __V = (V); \
+  (__m256)__builtin_ia32_roundps256((__v8sf)__V, (M)); })
+
+#define _mm256_ceil_pd(V)  _mm256_round_pd((V), _MM_FROUND_CEIL)
+#define _mm256_floor_pd(V) _mm256_round_pd((V), _MM_FROUND_FLOOR)
+#define _mm256_ceil_ps(V)  _mm256_round_ps((V), _MM_FROUND_CEIL)
+#define _mm256_floor_ps(V) _mm256_round_ps((V), _MM_FROUND_FLOOR)
+
+/* Logical */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_and_pd(__m256d a, __m256d b)
+{
+  return (__m256d)((__v4di)a & (__v4di)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_and_ps(__m256 a, __m256 b)
+{
+  return (__m256)((__v8si)a & (__v8si)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_pd(__m256d a, __m256d b)
+{
+  return (__m256d)(~(__v4di)a & (__v4di)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_ps(__m256 a, __m256 b)
+{
+  return (__m256)(~(__v8si)a & (__v8si)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_or_pd(__m256d a, __m256d b)
+{
+  return (__m256d)((__v4di)a | (__v4di)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_or_ps(__m256 a, __m256 b)
+{
+  return (__m256)((__v8si)a | (__v8si)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_pd(__m256d a, __m256d b)
+{
+  return (__m256d)((__v4di)a ^ (__v4di)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_ps(__m256 a, __m256 b)
+{
+  return (__m256)((__v8si)a ^ (__v8si)b);
+}
+
+/* Horizontal arithmetic */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_pd(__m256d a, __m256d b)
+{
+  return (__m256d)__builtin_ia32_haddpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_haddps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_pd(__m256d a, __m256d b)
+{
+  return (__m256d)__builtin_ia32_hsubpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_ps(__m256 a, __m256 b)
+{
+  return (__m256)__builtin_ia32_hsubps256((__v8sf)a, (__v8sf)b);
+}
+
+/* Vector permutations */
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_permutevar_pd(__m128d a, __m128i c)
+{
+  return (__m128d)__builtin_ia32_vpermilvarpd((__v2df)a, (__v2di)c);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar_pd(__m256d a, __m256i c)
+{
+  return (__m256d)__builtin_ia32_vpermilvarpd256((__v4df)a, (__v4di)c);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_permutevar_ps(__m128 a, __m128i c)
+{
+  return (__m128)__builtin_ia32_vpermilvarps((__v4sf)a, (__v4si)c);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar_ps(__m256 a, __m256i c)
+{
+  return (__m256)__builtin_ia32_vpermilvarps256((__v8sf)a,
+						  (__v8si)c);
+}
+
+#define _mm_permute_pd(A, C) __extension__ ({ \
+  __m128d __A = (A); \
+  (__m128d)__builtin_shufflevector((__v2df)__A, (__v2df) _mm_setzero_pd(), \
+                                   (C) & 0x1, ((C) & 0x2) >> 1); })
+
+#define _mm256_permute_pd(A, C) __extension__ ({ \
+  __m256d __A = (A); \
+  (__m256d)__builtin_shufflevector((__v4df)__A, (__v4df) _mm256_setzero_pd(), \
+                                   (C) & 0x1, ((C) & 0x2) >> 1, \
+                                   2 + (((C) & 0x4) >> 2), \
+                                   2 + (((C) & 0x8) >> 3)); })
+
+#define _mm_permute_ps(A, C) __extension__ ({ \
+  __m128 __A = (A); \
+  (__m128)__builtin_shufflevector((__v4sf)__A, (__v4sf) _mm_setzero_ps(), \
+                                   (C) & 0x3, ((C) & 0xc) >> 2, \
+                                   ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6); })
+
+#define _mm256_permute_ps(A, C) __extension__ ({ \
+  __m256 __A = (A); \
+  (__m256)__builtin_shufflevector((__v8sf)__A, (__v8sf) _mm256_setzero_ps(), \
+                                  (C) & 0x3, ((C) & 0xc) >> 2, \
+                                  ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6, \
+                                  4 + (((C) & 0x03) >> 0), \
+                                  4 + (((C) & 0x0c) >> 2), \
+                                  4 + (((C) & 0x30) >> 4), \
+                                  4 + (((C) & 0xc0) >> 6)); })
+
+#define _mm256_permute2f128_pd(V1, V2, M) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m256d __V2 = (V2); \
+  (__m256d)__builtin_ia32_vperm2f128_pd256((__v4df)__V1, (__v4df)__V2, (M)); })
+
+#define _mm256_permute2f128_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_ia32_vperm2f128_ps256((__v8sf)__V1, (__v8sf)__V2, (M)); })
+
+#define _mm256_permute2f128_si256(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_vperm2f128_si256((__v8si)__V1, (__v8si)__V2, (M)); })
+
+/* Vector Blend */
+#define _mm256_blend_pd(V1, V2, M) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m256d __V2 = (V2); \
+  (__m256d)__builtin_ia32_blendpd256((__v4df)__V1, (__v4df)__V2, (M)); })
+
+#define _mm256_blend_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_ia32_blendps256((__v8sf)__V1, (__v8sf)__V2, (M)); })
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_pd(__m256d a, __m256d b, __m256d c)
+{
+  return (__m256d)__builtin_ia32_blendvpd256((__v4df)a, (__v4df)b, (__v4df)c);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_ps(__m256 a, __m256 b, __m256 c)
+{
+  return (__m256)__builtin_ia32_blendvps256((__v8sf)a, (__v8sf)b, (__v8sf)c);
+}
+
+/* Vector Dot Product */
+#define _mm256_dp_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_ia32_dpps256((__v8sf)__V1, (__v8sf)__V2, (M)); })
+
+/* Vector shuffle */
+#define _mm256_shuffle_ps(a, b, mask) __extension__ ({ \
+        __m256 __a = (a); \
+        __m256 __b = (b); \
+        (__m256)__builtin_shufflevector((__v8sf)__a, (__v8sf)__b, \
+        (mask) & 0x3,                ((mask) & 0xc) >> 2, \
+        (((mask) & 0x30) >> 4) + 8,  (((mask) & 0xc0) >> 6) + 8, \
+        ((mask) & 0x3) + 4,          (((mask) & 0xc) >> 2) + 4, \
+        (((mask) & 0x30) >> 4) + 12, (((mask) & 0xc0) >> 6) + 12); })
+
+#define _mm256_shuffle_pd(a, b, mask) __extension__ ({ \
+        __m256d __a = (a); \
+        __m256d __b = (b); \
+        (__m256d)__builtin_shufflevector((__v4df)__a, (__v4df)__b, \
+        (mask) & 0x1, \
+        (((mask) & 0x2) >> 1) + 4, \
+        (((mask) & 0x4) >> 2) + 2, \
+        (((mask) & 0x8) >> 3) + 6); })
+
+/* Compare */
+#define _CMP_EQ_OQ    0x00 /* Equal (ordered, non-signaling)  */
+#define _CMP_LT_OS    0x01 /* Less-than (ordered, signaling)  */
+#define _CMP_LE_OS    0x02 /* Less-than-or-equal (ordered, signaling)  */
+#define _CMP_UNORD_Q  0x03 /* Unordered (non-signaling)  */
+#define _CMP_NEQ_UQ   0x04 /* Not-equal (unordered, non-signaling)  */
+#define _CMP_NLT_US   0x05 /* Not-less-than (unordered, signaling)  */
+#define _CMP_NLE_US   0x06 /* Not-less-than-or-equal (unordered, signaling)  */
+#define _CMP_ORD_Q    0x07 /* Ordered (nonsignaling)   */
+#define _CMP_EQ_UQ    0x08 /* Equal (unordered, non-signaling)  */
+#define _CMP_NGE_US   0x09 /* Not-greater-than-or-equal (unord, signaling)  */
+#define _CMP_NGT_US   0x0a /* Not-greater-than (unordered, signaling)  */
+#define _CMP_FALSE_OQ 0x0b /* False (ordered, non-signaling)  */
+#define _CMP_NEQ_OQ   0x0c /* Not-equal (ordered, non-signaling)  */
+#define _CMP_GE_OS    0x0d /* Greater-than-or-equal (ordered, signaling)  */
+#define _CMP_GT_OS    0x0e /* Greater-than (ordered, signaling)  */
+#define _CMP_TRUE_UQ  0x0f /* True (unordered, non-signaling)  */
+#define _CMP_EQ_OS    0x10 /* Equal (ordered, signaling)  */
+#define _CMP_LT_OQ    0x11 /* Less-than (ordered, non-signaling)  */
+#define _CMP_LE_OQ    0x12 /* Less-than-or-equal (ordered, non-signaling)  */
+#define _CMP_UNORD_S  0x13 /* Unordered (signaling)  */
+#define _CMP_NEQ_US   0x14 /* Not-equal (unordered, signaling)  */
+#define _CMP_NLT_UQ   0x15 /* Not-less-than (unordered, non-signaling)  */
+#define _CMP_NLE_UQ   0x16 /* Not-less-than-or-equal (unord, non-signaling)  */
+#define _CMP_ORD_S    0x17 /* Ordered (signaling)  */
+#define _CMP_EQ_US    0x18 /* Equal (unordered, signaling)  */
+#define _CMP_NGE_UQ   0x19 /* Not-greater-than-or-equal (unord, non-sign)  */
+#define _CMP_NGT_UQ   0x1a /* Not-greater-than (unordered, non-signaling)  */
+#define _CMP_FALSE_OS 0x1b /* False (ordered, signaling)  */
+#define _CMP_NEQ_OS   0x1c /* Not-equal (ordered, signaling)  */
+#define _CMP_GE_OQ    0x1d /* Greater-than-or-equal (ordered, non-signaling)  */
+#define _CMP_GT_OQ    0x1e /* Greater-than (ordered, non-signaling)  */
+#define _CMP_TRUE_US  0x1f /* True (unordered, signaling)  */
+
+#define _mm_cmp_pd(a, b, c) __extension__ ({ \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  (__m128d)__builtin_ia32_cmppd((__v2df)__a, (__v2df)__b, (c)); })
+
+#define _mm_cmp_ps(a, b, c) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_ia32_cmpps((__v4sf)__a, (__v4sf)__b, (c)); })
+
+#define _mm256_cmp_pd(a, b, c) __extension__ ({ \
+  __m256d __a = (a); \
+  __m256d __b = (b); \
+  (__m256d)__builtin_ia32_cmppd256((__v4df)__a, (__v4df)__b, (c)); })
+
+#define _mm256_cmp_ps(a, b, c) __extension__ ({ \
+  __m256 __a = (a); \
+  __m256 __b = (b); \
+  (__m256)__builtin_ia32_cmpps256((__v8sf)__a, (__v8sf)__b, (c)); })
+
+#define _mm_cmp_sd(a, b, c) __extension__ ({ \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  (__m128d)__builtin_ia32_cmpsd((__v2df)__a, (__v2df)__b, (c)); })
+
+#define _mm_cmp_ss(a, b, c) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_ia32_cmpss((__v4sf)__a, (__v4sf)__b, (c)); })
+
+/* Vector extract */
+#define _mm256_extractf128_pd(A, O) __extension__ ({ \
+  __m256d __A = (A); \
+  (__m128d)__builtin_ia32_vextractf128_pd256((__v4df)__A, (O)); })
+
+#define _mm256_extractf128_ps(A, O) __extension__ ({ \
+  __m256 __A = (A); \
+  (__m128)__builtin_ia32_vextractf128_ps256((__v8sf)__A, (O)); })
+
+#define _mm256_extractf128_si256(A, O) __extension__ ({ \
+  __m256i __A = (A); \
+  (__m128i)__builtin_ia32_vextractf128_si256((__v8si)__A, (O)); })
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi32(__m256i a, int const imm)
+{
+  __v8si b = (__v8si)a;
+  return b[imm];
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi16(__m256i a, int const imm)
+{
+  __v16hi b = (__v16hi)a;
+  return b[imm];
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi8(__m256i a, int const imm)
+{
+  __v32qi b = (__v32qi)a;
+  return b[imm];
+}
+
+#ifdef __x86_64__
+static __inline long long  __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi64(__m256i a, const int imm)
+{
+  __v4di b = (__v4di)a;
+  return b[imm];
+}
+#endif
+
+/* Vector insert */
+#define _mm256_insertf128_pd(V1, V2, O) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m128d __V2 = (V2); \
+  (__m256d)__builtin_ia32_vinsertf128_pd256((__v4df)__V1, (__v2df)__V2, (O)); })
+
+#define _mm256_insertf128_ps(V1, V2, O) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m128 __V2 = (V2); \
+  (__m256)__builtin_ia32_vinsertf128_ps256((__v8sf)__V1, (__v4sf)__V2, (O)); })
+
+#define _mm256_insertf128_si256(V1, V2, O) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m256i)__builtin_ia32_vinsertf128_si256((__v8si)__V1, (__v4si)__V2, (O)); })
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi32(__m256i a, int b, int const imm)
+{
+  __v8si c = (__v8si)a;
+  c[imm & 7] = b;
+  return (__m256i)c;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi16(__m256i a, int b, int const imm)
+{
+  __v16hi c = (__v16hi)a;
+  c[imm & 15] = b;
+  return (__m256i)c;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi8(__m256i a, int b, int const imm)
+{
+  __v32qi c = (__v32qi)a;
+  c[imm & 31] = b;
+  return (__m256i)c;
+}
+
+#ifdef __x86_64__
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi64(__m256i a, int b, int const imm)
+{
+  __v4di c = (__v4di)a;
+  c[imm & 3] = b;
+  return (__m256i)c;
+}
+#endif
+
+/* Conversion */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_pd(__m128i a)
+{
+  return (__m256d)__builtin_ia32_cvtdq2pd256((__v4si) a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_ps(__m256i a)
+{
+  return (__m256)__builtin_ia32_cvtdq2ps256((__v8si) a);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtpd_ps(__m256d a)
+{
+  return (__m128)__builtin_ia32_cvtpd2ps256((__v4df) a);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtps_epi32(__m256 a)
+{
+  return (__m256i)__builtin_ia32_cvtps2dq256((__v8sf) a);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtps_pd(__m128 a)
+{
+  return (__m256d)__builtin_ia32_cvtps2pd256((__v4sf) a);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvttpd_epi32(__m256d a)
+{
+  return (__m128i)__builtin_ia32_cvttpd2dq256((__v4df) a);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtpd_epi32(__m256d a)
+{
+  return (__m128i)__builtin_ia32_cvtpd2dq256((__v4df) a);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvttps_epi32(__m256 a)
+{
+  return (__m256i)__builtin_ia32_cvttps2dq256((__v8sf) a);
+}
+
+/* Vector replicate */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_movehdup_ps(__m256 a)
+{
+  return __builtin_shufflevector(a, a, 1, 1, 3, 3, 5, 5, 7, 7);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_moveldup_ps(__m256 a)
+{
+  return __builtin_shufflevector(a, a, 0, 0, 2, 2, 4, 4, 6, 6);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_movedup_pd(__m256d a)
+{
+  return __builtin_shufflevector(a, a, 0, 0, 2, 2);
+}
+
+/* Unpack and Interleave */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_pd(__m256d a, __m256d b)
+{
+  return __builtin_shufflevector(a, b, 1, 5, 1+2, 5+2);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_pd(__m256d a, __m256d b)
+{
+  return __builtin_shufflevector(a, b, 0, 4, 0+2, 4+2);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_ps(__m256 a, __m256 b)
+{
+  return __builtin_shufflevector(a, b, 2, 10, 2+1, 10+1, 6, 14, 6+1, 14+1);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_ps(__m256 a, __m256 b)
+{
+  return __builtin_shufflevector(a, b, 0, 8, 0+1, 8+1, 4, 12, 4+1, 12+1);
+}
+
+/* Bit Test */
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_vtestzpd((__v2df)a, (__v2df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_vtestcpd((__v2df)a, (__v2df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_vtestnzcpd((__v2df)a, (__v2df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_vtestzps((__v4sf)a, (__v4sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_vtestcps((__v4sf)a, (__v4sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_vtestnzcps((__v4sf)a, (__v4sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_pd(__m256d a, __m256d b)
+{
+  return __builtin_ia32_vtestzpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_pd(__m256d a, __m256d b)
+{
+  return __builtin_ia32_vtestcpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_pd(__m256d a, __m256d b)
+{
+  return __builtin_ia32_vtestnzcpd256((__v4df)a, (__v4df)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_ps(__m256 a, __m256 b)
+{
+  return __builtin_ia32_vtestzps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_ps(__m256 a, __m256 b)
+{
+  return __builtin_ia32_vtestcps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_ps(__m256 a, __m256 b)
+{
+  return __builtin_ia32_vtestnzcps256((__v8sf)a, (__v8sf)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_si256(__m256i a, __m256i b)
+{
+  return __builtin_ia32_ptestz256((__v4di)a, (__v4di)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_si256(__m256i a, __m256i b)
+{
+  return __builtin_ia32_ptestc256((__v4di)a, (__v4di)b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_si256(__m256i a, __m256i b)
+{
+  return __builtin_ia32_ptestnzc256((__v4di)a, (__v4di)b);
+}
+
+/* Vector extract sign mask */
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_pd(__m256d a)
+{
+  return __builtin_ia32_movmskpd256((__v4df)a);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_ps(__m256 a)
+{
+  return __builtin_ia32_movmskps256((__v8sf)a);
+}
+
+/* Vector zero */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_zeroall(void)
+{
+  __builtin_ia32_vzeroall();
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_zeroupper(void)
+{
+  __builtin_ia32_vzeroupper();
+}
+
+/* Vector load with broadcast */
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_broadcast_ss(float const *a)
+{
+  return (__m128)__builtin_ia32_vbroadcastss(a);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_sd(double const *a)
+{
+  return (__m256d)__builtin_ia32_vbroadcastsd256(a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_ss(float const *a)
+{
+  return (__m256)__builtin_ia32_vbroadcastss256(a);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_pd(__m128d const *a)
+{
+  return (__m256d)__builtin_ia32_vbroadcastf128_pd256(a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_ps(__m128 const *a)
+{
+  return (__m256)__builtin_ia32_vbroadcastf128_ps256(a);
+}
+
+/* SIMD load ops */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_load_pd(double const *p)
+{
+  return *(__m256d *)p;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_load_ps(float const *p)
+{
+  return *(__m256 *)p;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_pd(double const *p)
+{
+  struct __loadu_pd {
+    __m256d v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_pd*)p)->v;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_ps(float const *p)
+{
+  struct __loadu_ps {
+    __m256 v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_ps*)p)->v;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_load_si256(__m256i const *p)
+{
+  return *p;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_si256(__m256i const *p)
+{
+  struct __loadu_si256 {
+    __m256i v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_si256*)p)->v;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_lddqu_si256(__m256i const *p)
+{
+  return (__m256i)__builtin_ia32_lddqu256((char const *)p);
+}
+
+/* SIMD store ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_pd(double *p, __m256d a)
+{
+  *(__m256d *)p = a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_ps(float *p, __m256 a)
+{
+  *(__m256 *)p = a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_pd(double *p, __m256d a)
+{
+  __builtin_ia32_storeupd256(p, (__v4df)a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_ps(float *p, __m256 a)
+{
+  __builtin_ia32_storeups256(p, (__v8sf)a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_si256(__m256i *p, __m256i a)
+{
+  *p = a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_si256(__m256i *p, __m256i a)
+{
+  __builtin_ia32_storedqu256((char *)p, (__v32qi)a);
+}
+
+/* Conditional load ops */
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_pd(double const *p, __m128d m)
+{
+  return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)p, (__v2df)m);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_pd(double const *p, __m256d m)
+{
+  return (__m256d)__builtin_ia32_maskloadpd256((const __v4df *)p, (__v4df)m);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_ps(float const *p, __m128 m)
+{
+  return (__m128)__builtin_ia32_maskloadps((const __v4sf *)p, (__v4sf)m);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_ps(float const *p, __m256 m)
+{
+  return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)p, (__v8sf)m);
+}
+
+/* Conditional store ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_ps(float *p, __m256 m, __m256 a)
+{
+  __builtin_ia32_maskstoreps256((__v8sf *)p, (__v8sf)m, (__v8sf)a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_pd(double *p, __m128d m, __m128d a)
+{
+  __builtin_ia32_maskstorepd((__v2df *)p, (__v2df)m, (__v2df)a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_pd(double *p, __m256d m, __m256d a)
+{
+  __builtin_ia32_maskstorepd256((__v4df *)p, (__v4df)m, (__v4df)a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_ps(float *p, __m128 m, __m128 a)
+{
+  __builtin_ia32_maskstoreps((__v4sf *)p, (__v4sf)m, (__v4sf)a);
+}
+
+/* Cacheability support ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_si256(__m256i *a, __m256i b)
+{
+  __builtin_ia32_movntdq256((__v4di *)a, (__v4di)b);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_pd(double *a, __m256d b)
+{
+  __builtin_ia32_movntpd256(a, (__v4df)b);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_ps(float *p, __m256 a)
+{
+  __builtin_ia32_movntps256(p, (__v8sf)a);
+}
+
+/* Create vectors */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_set_pd(double a, double b, double c, double d)
+{
+  return (__m256d){ d, c, b, a };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_set_ps(float a, float b, float c, float d,
+	            float e, float f, float g, float h)
+{
+  return (__m256){ h, g, f, e, d, c, b, a };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi32(int i0, int i1, int i2, int i3,
+		             int i4, int i5, int i6, int i7)
+{
+  return (__m256i)(__v8si){ i7, i6, i5, i4, i3, i2, i1, i0 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi16(short w15, short w14, short w13, short w12,
+		             short w11, short w10, short w09, short w08,
+		             short w07, short w06, short w05, short w04,
+		             short w03, short w02, short w01, short w00)
+{
+  return (__m256i)(__v16hi){ w00, w01, w02, w03, w04, w05, w06, w07,
+                             w08, w09, w10, w11, w12, w13, w14, w15 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi8(char b31, char b30, char b29, char b28,
+		            char b27, char b26, char b25, char b24,
+		            char b23, char b22, char b21, char b20,
+		            char b19, char b18, char b17, char b16,
+		            char b15, char b14, char b13, char b12,
+		            char b11, char b10, char b09, char b08,
+		            char b07, char b06, char b05, char b04,
+		            char b03, char b02, char b01, char b00)
+{
+  return (__m256i)(__v32qi){
+    b00, b01, b02, b03, b04, b05, b06, b07,
+    b08, b09, b10, b11, b12, b13, b14, b15,
+    b16, b17, b18, b19, b20, b21, b22, b23,
+    b24, b25, b26, b27, b28, b29, b30, b31
+  };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi64x(long long a, long long b, long long c, long long d)
+{
+  return (__m256i)(__v4di){ d, c, b, a };
+}
+
+/* Create vectors with elements in reverse order */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_pd(double a, double b, double c, double d)
+{
+  return (__m256d){ a, b, c, d };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_ps(float a, float b, float c, float d,
+		           float e, float f, float g, float h)
+{
+  return (__m256){ a, b, c, d, e, f, g, h };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi32(int i0, int i1, int i2, int i3,
+		              int i4, int i5, int i6, int i7)
+{
+  return (__m256i)(__v8si){ i0, i1, i2, i3, i4, i5, i6, i7 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi16(short w15, short w14, short w13, short w12,
+		   short w11, short w10, short w09, short w08,
+		   short w07, short w06, short w05, short w04,
+		   short w03, short w02, short w01, short w00)
+{
+  return (__m256i)(__v16hi){ w15, w14, w13, w12, w11, w10, w09, w08,
+			                       w07, w06, w05, w04, w03, w02, w01, w00 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi8(char b31, char b30, char b29, char b28,
+		             char b27, char b26, char b25, char b24,
+		             char b23, char b22, char b21, char b20,
+		             char b19, char b18, char b17, char b16,
+		             char b15, char b14, char b13, char b12,
+		             char b11, char b10, char b09, char b08,
+		             char b07, char b06, char b05, char b04,
+		             char b03, char b02, char b01, char b00)
+{
+  return (__m256i)(__v32qi){
+    b31, b30, b29, b28, b27, b26, b25, b24,
+		b23, b22, b21, b20, b19, b18, b17, b16,
+		b15, b14, b13, b12, b11, b10, b09, b08,
+		b07, b06, b05, b04, b03, b02, b01, b00 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi64x(long long a, long long b, long long c, long long d)
+{
+  return (__m256i)(__v4di){ a, b, c, d };
+}
+
+/* Create vectors with repeated elements */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_pd(double w)
+{
+  return (__m256d){ w, w, w, w };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_ps(float w)
+{
+  return (__m256){ w, w, w, w, w, w, w, w };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi32(int i)
+{
+  return (__m256i)(__v8si){ i, i, i, i, i, i, i, i };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi16(short w)
+{
+  return (__m256i)(__v16hi){ w, w, w, w, w, w, w, w, w, w, w, w, w, w, w, w };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi8(char b)
+{
+  return (__m256i)(__v32qi){ b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, b,
+                             b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, b };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi64x(long long q)
+{
+  return (__m256i)(__v4di){ q, q, q, q };
+}
+
+/* Create zeroed vectors */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_pd(void)
+{
+  return (__m256d){ 0, 0, 0, 0 };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_ps(void)
+{
+  return (__m256){ 0, 0, 0, 0, 0, 0, 0, 0 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_si256(void)
+{
+  return (__m256i){ 0LL, 0LL, 0LL, 0LL };
+}
+
+/* Cast between vector types */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd_ps(__m256d in)
+{
+  return (__m256)in;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd_si256(__m256d in)
+{
+  return (__m256i)in;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castps_pd(__m256 in)
+{
+  return (__m256d)in;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castps_si256(__m256 in)
+{
+  return (__m256i)in;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_ps(__m256i in)
+{
+  return (__m256)in;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_pd(__m256i in)
+{
+  return (__m256d)in;
+}
+
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd256_pd128(__m256d in)
+{
+  return __builtin_shufflevector(in, in, 0, 1);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm256_castps256_ps128(__m256 in)
+{
+  return __builtin_shufflevector(in, in, 0, 1, 2, 3);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_si128(__m256i in)
+{
+  return __builtin_shufflevector(in, in, 0, 1);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd128_pd256(__m128d in)
+{
+  __m128d zero = _mm_setzero_pd();
+  return __builtin_shufflevector(in, zero, 0, 1, 2, 2);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castps128_ps256(__m128 in)
+{
+  __m128 zero = _mm_setzero_ps();
+  return __builtin_shufflevector(in, zero, 0, 1, 2, 3, 4, 4, 4, 4);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi128_si256(__m128i in)
+{
+  __m128i zero = _mm_setzero_si128();
+  return __builtin_shufflevector(in, zero, 0, 1, 2, 2);
+}
+
+/* SIMD load ops (unaligned) */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128(float const *addr_hi, float const *addr_lo)
+{
+  struct __loadu_ps {
+    __m128 v;
+  } __attribute__((__packed__, __may_alias__));
+
+  __m256 v256 = _mm256_castps128_ps256(((struct __loadu_ps*)addr_lo)->v);
+  return _mm256_insertf128_ps(v256, ((struct __loadu_ps*)addr_hi)->v, 1);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128d(double const *addr_hi, double const *addr_lo)
+{
+  struct __loadu_pd {
+    __m128d v;
+  } __attribute__((__packed__, __may_alias__));
+  
+  __m256d v256 = _mm256_castpd128_pd256(((struct __loadu_pd*)addr_lo)->v);
+  return _mm256_insertf128_pd(v256, ((struct __loadu_pd*)addr_hi)->v, 1);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128i(__m128i const *addr_hi, __m128i const *addr_lo)
+{
+  struct __loadu_si128 {
+    __m128i v;
+  } __attribute__((packed, may_alias));
+  __m256i v256 = _mm256_castsi128_si256(((struct __loadu_si128*)addr_lo)->v);
+  return _mm256_insertf128_si256(v256, ((struct __loadu_si128*)addr_hi)->v, 1);
+}
+
+/* SIMD store ops (unaligned) */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128(float *addr_hi, float *addr_lo, __m256 a)
+{
+  __m128 v128;
+
+  v128 = _mm256_castps256_ps128(a);
+  __builtin_ia32_storeups(addr_lo, v128);
+  v128 = _mm256_extractf128_ps(a, 1);
+  __builtin_ia32_storeups(addr_hi, v128);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128d(double *addr_hi, double *addr_lo, __m256d a)
+{
+  __m128d v128;
+
+  v128 = _mm256_castpd256_pd128(a);
+  __builtin_ia32_storeupd(addr_lo, v128);
+  v128 = _mm256_extractf128_pd(a, 1);
+  __builtin_ia32_storeupd(addr_hi, v128);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128i(__m128i *addr_hi, __m128i *addr_lo, __m256i a)
+{
+  __m128i v128;
+
+  v128 = _mm256_castsi256_si128(a);
+  __builtin_ia32_storedqu((char *)addr_lo, (__v16qi)v128);
+  v128 = _mm256_extractf128_si256(a, 1);
+  __builtin_ia32_storedqu((char *)addr_hi, (__v16qi)v128);
+}
diff --git a/clang/lib/Headers/bmi2intrin.h b/clang/lib/Headers/bmi2intrin.h
new file mode 100644
index 0000000..c60b0c4
--- /dev/null
+++ b/clang/lib/Headers/bmi2intrin.h
@@ -0,0 +1,75 @@
+/*===---- bmi2intrin.h - BMI2 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <bmi2intrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __BMI2__
+# error "BMI2 instruction set not enabled"
+#endif /* __BMI2__ */
+
+#ifndef __BMI2INTRIN_H
+#define __BMI2INTRIN_H
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_bzhi_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_bzhi_si(__X, __Y);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_pdep_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_pdep_si(__X, __Y);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_pext_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_pext_si(__X, __Y);
+}
+
+#ifdef  __x86_64__
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_bzhi_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_bzhi_di(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_pdep_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_pdep_di(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_pext_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_pext_di(__X, __Y);
+}
+
+#endif /* !__x86_64__  */
+
+#endif /* __BMI2INTRIN_H */
diff --git a/clang/lib/Headers/bmiintrin.h b/clang/lib/Headers/bmiintrin.h
new file mode 100644
index 0000000..2f7db73
--- /dev/null
+++ b/clang/lib/Headers/bmiintrin.h
@@ -0,0 +1,115 @@
+/*===---- bmiintrin.h - BMI intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <bmiintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __BMI__
+# error "BMI instruction set not enabled"
+#endif /* __BMI__ */
+
+#ifndef __BMIINTRIN_H
+#define __BMIINTRIN_H
+
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__tzcnt16(unsigned short __X)
+{
+  return __builtin_ctzs(__X);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__andn_u32(unsigned int __X, unsigned int __Y)
+{
+  return ~__X & __Y;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__bextr_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_bextr_u32(__X, __Y);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsi_u32(unsigned int __X)
+{
+  return __X & -__X;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsmsk_u32(unsigned int __X)
+{
+  return __X ^ (__X - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsr_u32(unsigned int __X)
+{
+  return __X & (__X - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__tzcnt32(unsigned int __X)
+{
+  return __builtin_ctz(__X);
+}
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__andn_u64 (unsigned long long __X, unsigned long long __Y)
+{
+  return ~__X & __Y;
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__bextr_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_bextr_u64(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsi_u64(unsigned long long __X)
+{
+  return __X & -__X;
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsmsk_u64(unsigned long long __X)
+{
+  return __X ^ (__X - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsr_u64(unsigned long long __X)
+{
+  return __X & (__X - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__tzcnt64(unsigned long long __X)
+{
+  return __builtin_ctzll(__X);
+}
+#endif
+
+#endif /* __BMIINTRIN_H */
diff --git a/clang/lib/Headers/cpuid.h b/clang/lib/Headers/cpuid.h
new file mode 100644
index 0000000..05c293f
--- /dev/null
+++ b/clang/lib/Headers/cpuid.h
@@ -0,0 +1,33 @@
+/*===---- cpuid.h - X86 cpu model detection --------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !(__x86_64__ || __i386__)
+#error this header is for x86 only
+#endif
+
+static inline int __get_cpuid (unsigned int level, unsigned int *eax,
+                               unsigned int *ebx, unsigned int *ecx,
+                               unsigned int *edx) {
+    asm("cpuid" : "=a"(*eax), "=b" (*ebx), "=c"(*ecx), "=d"(*edx) : "0"(level));
+    return 1;
+}
diff --git a/clang/lib/Headers/emmintrin.h b/clang/lib/Headers/emmintrin.h
new file mode 100644
index 0000000..e10b77d
--- /dev/null
+++ b/clang/lib/Headers/emmintrin.h
@@ -0,0 +1,1424 @@
+/*===---- emmintrin.h - SSE2 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __EMMINTRIN_H
+#define __EMMINTRIN_H
+
+#ifndef __SSE2__
+#error "SSE2 instruction set not enabled"
+#else
+
+#include <xmmintrin.h>
+
+typedef double __m128d __attribute__((__vector_size__(16)));
+typedef long long __m128i __attribute__((__vector_size__(16)));
+
+/* Type defines.  */
+typedef double __v2df __attribute__ ((__vector_size__ (16)));
+typedef long long __v2di __attribute__ ((__vector_size__ (16)));
+typedef short __v8hi __attribute__((__vector_size__(16)));
+typedef char __v16qi __attribute__((__vector_size__(16)));
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_add_sd(__m128d a, __m128d b)
+{
+  a[0] += b[0];
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_add_pd(__m128d a, __m128d b)
+{
+  return a + b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sub_sd(__m128d a, __m128d b)
+{
+  a[0] -= b[0];
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pd(__m128d a, __m128d b)
+{
+  return a - b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_mul_sd(__m128d a, __m128d b)
+{
+  a[0] *= b[0];
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_mul_pd(__m128d a, __m128d b)
+{
+  return a * b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_div_sd(__m128d a, __m128d b)
+{
+  a[0] /= b[0];
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_div_pd(__m128d a, __m128d b)
+{
+  return a / b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_sd(__m128d a, __m128d b)
+{
+  __m128d c = __builtin_ia32_sqrtsd(b);
+  return (__m128d) { c[0], a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_pd(__m128d a)
+{
+  return __builtin_ia32_sqrtpd(a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_min_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_minsd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_min_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_minpd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_max_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_maxsd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_max_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_maxpd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_and_pd(__m128d a, __m128d b)
+{
+  return (__m128d)((__v4si)a & (__v4si)b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_pd(__m128d a, __m128d b)
+{
+  return (__m128d)(~(__v4si)a & (__v4si)b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_or_pd(__m128d a, __m128d b)
+{
+  return (__m128d)((__v4si)a | (__v4si)b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_xor_pd(__m128d a, __m128d b)
+{
+  return (__m128d)((__v4si)a ^ (__v4si)b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(b, a, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(b, a, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 7);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 3);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 4);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(a, b, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(b, a, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_pd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmppd(b, a, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(b, a, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(b, a, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 7);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 3);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 4);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(a, b, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(b, a, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_sd(__m128d a, __m128d b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(b, a, 6);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comieq_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdeq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comilt_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdlt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comile_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdle(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comigt_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdgt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comige_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdge(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comineq_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_comisdneq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomieq_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdeq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomilt_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdlt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomile_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdle(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomigt_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdgt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomige_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdge(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomineq_sd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_ucomisdneq(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_ps(__m128d a)
+{
+  return __builtin_ia32_cvtpd2ps(a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pd(__m128 a)
+{
+  return __builtin_ia32_cvtps2pd(a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_pd(__m128i a)
+{
+  return __builtin_ia32_cvtdq2pd((__v4si)a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_epi32(__m128d a)
+{
+  return __builtin_ia32_cvtpd2dq(a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_si32(__m128d a)
+{
+  return __builtin_ia32_cvtsd2si(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_ss(__m128 a, __m128d b)
+{
+  a[0] = b[0];
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_sd(__m128d a, int b)
+{
+  a[0] = b;
+  return a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_sd(__m128d a, __m128 b)
+{
+  a[0] = b[0];
+  return a;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvttpd_epi32(__m128d a)
+{
+  return (__m128i)__builtin_ia32_cvttpd2dq(a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvttsd_si32(__m128d a)
+{
+  return a[0];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_pi32(__m128d a)
+{
+  return (__m64)__builtin_ia32_cvtpd2pi(a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvttpd_pi32(__m128d a)
+{
+  return (__m64)__builtin_ia32_cvttpd2pi(a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32_pd(__m64 a)
+{
+  return __builtin_ia32_cvtpi2pd((__v2si)a);
+}
+
+static __inline__ double __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_f64(__m128d a)
+{
+  return a[0];
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load_pd(double const *dp)
+{
+  return *(__m128d*)dp;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load1_pd(double const *dp)
+{
+  struct __mm_load1_pd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  double u = ((struct __mm_load1_pd_struct*)dp)->u;
+  return (__m128d){ u, u };
+}
+
+#define        _mm_load_pd1(dp)        _mm_load1_pd(dp)
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadr_pd(double const *dp)
+{
+  __m128d u = *(__m128d*)dp;
+  return __builtin_shufflevector(u, u, 1, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_pd(double const *dp)
+{
+  struct __loadu_pd {
+    __m128d v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_pd*)dp)->v;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load_sd(double const *dp)
+{
+  struct __mm_load_sd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  double u = ((struct __mm_load_sd_struct*)dp)->u;
+  return (__m128d){ u, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadh_pd(__m128d a, double const *dp)
+{
+  struct __mm_loadh_pd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  double u = ((struct __mm_loadh_pd_struct*)dp)->u;
+  return (__m128d){ a[0], u };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_pd(__m128d a, double const *dp)
+{
+  struct __mm_loadl_pd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  double u = ((struct __mm_loadl_pd_struct*)dp)->u;
+  return (__m128d){ u, a[1] }; 
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set_sd(double w)
+{
+  return (__m128d){ w, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pd(double w)
+{
+  return (__m128d){ w, w };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set_pd(double w, double x)
+{
+  return (__m128d){ x, w };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pd(double w, double x)
+{
+  return (__m128d){ w, x };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_pd(void)
+{
+  return (__m128d){ 0, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_move_sd(__m128d a, __m128d b)
+{
+  return (__m128d){ b[0], a[1] };
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_sd(double *dp, __m128d a)
+{
+  struct __mm_store_sd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store_sd_struct*)dp)->u = a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store1_pd(double *dp, __m128d a)
+{
+  struct __mm_store1_pd_struct {
+    double u[2];
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store1_pd_struct*)dp)->u[0] = a[0];
+  ((struct __mm_store1_pd_struct*)dp)->u[1] = a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_pd(double *dp, __m128d a)
+{
+  *(__m128d *)dp = a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_pd(double *dp, __m128d a)
+{
+  __builtin_ia32_storeupd(dp, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storer_pd(double *dp, __m128d a)
+{
+  a = __builtin_shufflevector(a, a, 1, 0);
+  *(__m128d *)dp = a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeh_pd(double *dp, __m128d a)
+{
+  struct __mm_storeh_pd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pd_struct*)dp)->u = a[1];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storel_pd(double *dp, __m128d a)
+{
+  struct __mm_storeh_pd_struct {
+    double u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pd_struct*)dp)->u = a[0];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)((__v16qi)a + (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)((__v8hi)a + (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)((__v4si)a + (__v4si)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_si64(__m64 a, __m64 b)
+{
+  return a + b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi64(__m128i a, __m128i b)
+{
+  return a + b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_paddsb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_paddsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epu8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_paddusb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epu16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_paddusw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_avg_epu8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pavgb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_avg_epu16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pavgw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_madd_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pmaddwd128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pmaxsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pmaxub128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pminsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pminub128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pmulhw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_epu16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_pmulhuw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)((__v8hi)a * (__v8hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_su32(__m64 a, __m64 b)
+{
+  return __builtin_ia32_pmuludq((__v2si)a, (__v2si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mul_epu32(__m128i a, __m128i b)
+{
+  return __builtin_ia32_pmuludq128((__v4si)a, (__v4si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sad_epu8(__m128i a, __m128i b)
+{
+  return __builtin_ia32_psadbw128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)((__v16qi)a - (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)((__v8hi)a - (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)((__v4si)a - (__v4si)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_si64(__m64 a, __m64 b)
+{
+  return a - b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi64(__m128i a, __m128i b)
+{
+  return a - b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_psubsb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_psubsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epu8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_psubusb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epu16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_psubusw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_and_si128(__m128i a, __m128i b)
+{
+  return a & b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_si128(__m128i a, __m128i b)
+{
+  return ~a & b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_or_si128(__m128i a, __m128i b)
+{
+  return a | b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_xor_si128(__m128i a, __m128i b)
+{
+  return a ^ b;
+}
+
+#define _mm_slli_si128(a, count) __extension__ ({ \
+  __m128i __a = (a); \
+  (__m128i)__builtin_ia32_pslldqi128(__a, (count)*8); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi16(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_psllwi128((__v8hi)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi16(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_psllw128((__v8hi)a, (__v8hi)count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi32(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_pslldi128((__v4si)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi32(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_pslld128((__v4si)a, (__v4si)count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi64(__m128i a, int count)
+{
+  return __builtin_ia32_psllqi128(a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi64(__m128i a, __m128i count)
+{
+  return __builtin_ia32_psllq128(a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srai_epi16(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_psrawi128((__v8hi)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sra_epi16(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_psraw128((__v8hi)a, (__v8hi)count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srai_epi32(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_psradi128((__v4si)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sra_epi32(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_psrad128((__v4si)a, (__v4si)count);
+}
+
+
+#define _mm_srli_si128(a, count) __extension__ ({ \
+  __m128i __a = (a); \
+  (__m128i)__builtin_ia32_psrldqi128(__a, (count)*8); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi16(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_psrlwi128((__v8hi)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi16(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_psrlw128((__v8hi)a, (__v8hi)count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi32(__m128i a, int count)
+{
+  return (__m128i)__builtin_ia32_psrldi128((__v4si)a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi32(__m128i a, __m128i count)
+{
+  return (__m128i)__builtin_ia32_psrld128((__v4si)a, (__v4si)count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi64(__m128i a, int count)
+{
+  return __builtin_ia32_psrlqi128(a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi64(__m128i a, __m128i count)
+{
+  return __builtin_ia32_psrlq128(a, count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)((__v16qi)a == (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)((__v8hi)a == (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)((__v4si)a == (__v4si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi8(__m128i a, __m128i b)
+{
+  /* This function always performs a signed comparison, but __v16qi is a char
+     which may be signed or unsigned. */
+  typedef signed char __v16qs __attribute__((__vector_size__(16)));
+  return (__m128i)((__v16qs)a > (__v16qs)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)((__v8hi)a > (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)((__v4si)a > (__v4si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi8(__m128i a, __m128i b)
+{
+  return _mm_cmpgt_epi8(b,a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi16(__m128i a, __m128i b)
+{
+  return _mm_cmpgt_epi16(b,a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi32(__m128i a, __m128i b)
+{
+  return _mm_cmpgt_epi32(b,a);
+}
+
+#ifdef __x86_64__
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_sd(__m128d a, long long b)
+{
+  a[0] = b;
+  return a;
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_si64(__m128d a)
+{
+  return __builtin_ia32_cvtsd2si64(a);
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvttsd_si64(__m128d a)
+{
+  return a[0];
+}
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_ps(__m128i a)
+{
+  return __builtin_ia32_cvtdq2ps((__v4si)a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_epi32(__m128 a)
+{
+  return (__m128i)__builtin_ia32_cvtps2dq(a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvttps_epi32(__m128 a)
+{
+  return (__m128i)__builtin_ia32_cvttps2dq(a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_si128(int a)
+{
+  return (__m128i)(__v4si){ a, 0, 0, 0 };
+}
+
+#ifdef __x86_64__
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_si128(long long a)
+{
+  return (__m128i){ a, 0 };
+}
+#endif
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi128_si32(__m128i a)
+{
+  __v4si b = (__v4si)a;
+  return b[0];
+}
+
+#ifdef __x86_64__
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi128_si64(__m128i a)
+{
+  return a[0];
+}
+#endif
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_load_si128(__m128i const *p)
+{
+  return *p;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_si128(__m128i const *p)
+{
+  struct __loadu_si128 {
+    __m128i v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_si128*)p)->v;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_epi64(__m128i const *p)
+{
+  struct __mm_loadl_epi64_struct {
+    long long u;
+  } __attribute__((__packed__, __may_alias__));
+  return (__m128i) { ((struct __mm_loadl_epi64_struct*)p)->u, 0};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi64x(long long q1, long long q0)
+{
+  return (__m128i){ q0, q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi64(__m64 q1, __m64 q0)
+{
+  return (__m128i){ (long long)q0, (long long)q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi32(int i3, int i2, int i1, int i0)
+{
+  return (__m128i)(__v4si){ i0, i1, i2, i3};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi16(short w7, short w6, short w5, short w4, short w3, short w2, short w1, short w0)
+{
+  return (__m128i)(__v8hi){ w0, w1, w2, w3, w4, w5, w6, w7 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi8(char b15, char b14, char b13, char b12, char b11, char b10, char b9, char b8, char b7, char b6, char b5, char b4, char b3, char b2, char b1, char b0)
+{
+  return (__m128i)(__v16qi){ b0, b1, b2, b3, b4, b5, b6, b7, b8, b9, b10, b11, b12, b13, b14, b15 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi64x(long long q)
+{
+  return (__m128i){ q, q };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi64(__m64 q)
+{
+  return (__m128i){ (long long)q, (long long)q };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi32(int i)
+{
+  return (__m128i)(__v4si){ i, i, i, i };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi16(short w)
+{
+  return (__m128i)(__v8hi){ w, w, w, w, w, w, w, w };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi8(char b)
+{
+  return (__m128i)(__v16qi){ b, b, b, b, b, b, b, b, b, b, b, b, b, b, b, b };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi64(__m64 q0, __m64 q1)
+{
+  return (__m128i){ (long long)q0, (long long)q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi32(int i0, int i1, int i2, int i3)
+{
+  return (__m128i)(__v4si){ i0, i1, i2, i3};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi16(short w0, short w1, short w2, short w3, short w4, short w5, short w6, short w7)
+{
+  return (__m128i)(__v8hi){ w0, w1, w2, w3, w4, w5, w6, w7 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi8(char b0, char b1, char b2, char b3, char b4, char b5, char b6, char b7, char b8, char b9, char b10, char b11, char b12, char b13, char b14, char b15)
+{
+  return (__m128i)(__v16qi){ b0, b1, b2, b3, b4, b5, b6, b7, b8, b9, b10, b11, b12, b13, b14, b15 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_si128(void)
+{
+  return (__m128i){ 0LL, 0LL };
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_si128(__m128i *p, __m128i b)
+{
+  *p = b;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_si128(__m128i *p, __m128i b)
+{
+  __builtin_ia32_storedqu((char *)p, (__v16qi)b);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskmoveu_si128(__m128i d, __m128i n, char *p)
+{
+  __builtin_ia32_maskmovdqu((__v16qi)d, (__v16qi)n, p);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storel_epi64(__m128i *p, __m128i a)
+{
+  __builtin_ia32_storelv4si((__v2si *)p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_pd(double *p, __m128d a)
+{
+  __builtin_ia32_movntpd(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_si128(__m128i *p, __m128i a)
+{
+  __builtin_ia32_movntdq(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_si32(int *p, int a)
+{
+  __builtin_ia32_movnti(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_clflush(void const *p)
+{
+  __builtin_ia32_clflush(p);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_lfence(void)
+{
+  __builtin_ia32_lfence();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_mfence(void)
+{
+  __builtin_ia32_mfence();
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packs_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_packsswb128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packs_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_packssdw128((__v4si)a, (__v4si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packus_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_ia32_packuswb128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_extract_epi16(__m128i a, int imm)
+{
+  __v8hi b = (__v8hi)a;
+  return (unsigned short)b[imm];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_insert_epi16(__m128i a, int b, int imm)
+{
+  __v8hi c = (__v8hi)a;
+  c[imm & 7] = b;
+  return (__m128i)c;
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_epi8(__m128i a)
+{
+  return __builtin_ia32_pmovmskb128((__v16qi)a);
+}
+
+#define _mm_shuffle_epi32(a, imm) __extension__ ({ \
+  __m128i __a = (a); \
+  (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si) _mm_set1_epi32(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6); })
+
+#define _mm_shufflelo_epi16(a, imm) __extension__ ({ \
+  __m128i __a = (a); \
+  (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi) _mm_set1_epi16(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4, 5, 6, 7); })
+
+#define _mm_shufflehi_epi16(a, imm) __extension__ ({ \
+  __m128i __a = (a); \
+  (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi) _mm_set1_epi16(0), \
+                                   0, 1, 2, 3, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v16qi)a, (__v16qi)b, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v8hi)a, (__v8hi)b, 4, 8+4, 5, 8+5, 6, 8+6, 7, 8+7);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v4si)a, (__v4si)b, 2, 4+2, 3, 4+3);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi64(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector(a, b, 1, 2+1);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi8(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v16qi)a, (__v16qi)b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi16(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v8hi)a, (__v8hi)b, 0, 8+0, 1, 8+1, 2, 8+2, 3, 8+3);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi32(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector((__v4si)a, (__v4si)b, 0, 4+0, 1, 4+1);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi64(__m128i a, __m128i b)
+{
+  return (__m128i)__builtin_shufflevector(a, b, 0, 2+0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_movepi64_pi64(__m128i a)
+{
+  return (__m64)a[0];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_movpi64_pi64(__m64 a)
+{
+  return (__m128i){ (long long)a, 0 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_move_epi64(__m128i a)
+{
+  return __builtin_shufflevector(a, (__m128i){ 0 }, 0, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pd(__m128d a, __m128d b)
+{
+  return __builtin_shufflevector(a, b, 1, 2+1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pd(__m128d a, __m128d b)
+{
+  return __builtin_shufflevector(a, b, 0, 2+0);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_pd(__m128d a)
+{
+  return __builtin_ia32_movmskpd(a);
+}
+
+#define _mm_shuffle_pd(a, b, i) __extension__ ({ \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  __builtin_shufflevector(__a, __b, (i) & 1, (((i) & 2) >> 1) + 2); })
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_castpd_ps(__m128d in)
+{
+  return (__m128)in;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_castpd_si128(__m128d in)
+{
+  return (__m128i)in;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_castps_pd(__m128 in)
+{
+  return (__m128d)in;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_castps_si128(__m128 in)
+{
+  return (__m128i)in;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_castsi128_ps(__m128i in)
+{
+  return (__m128)in;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_castsi128_pd(__m128i in)
+{
+  return (__m128d)in;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_pause(void)
+{
+  __asm__ volatile ("pause");
+}
+
+#define _MM_SHUFFLE2(x, y) (((x) << 1) | (y))
+
+#endif /* __SSE2__ */
+
+#endif /* __EMMINTRIN_H */
diff --git a/clang/lib/Headers/float.h b/clang/lib/Headers/float.h
new file mode 100644
index 0000000..65b517d
--- /dev/null
+++ b/clang/lib/Headers/float.h
@@ -0,0 +1,124 @@
+/*===---- float.h - Characteristics of floating point types ----------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __FLOAT_H
+#define __FLOAT_H
+
+/* If we're on MinGW, fall back to the system's float.h, which might have
+ * additional definitions provided for Windows.
+ * For more details see http://msdn.microsoft.com/en-us/library/y0ybw9fy.aspx
+ */
+#if defined(__MINGW32__) && \
+    defined(__has_include_next) && __has_include_next(<float.h>)
+#  include_next <float.h>
+
+/* Undefine anything that we'll be redefining below. */
+#  undef FLT_EVAL_METHOD
+#  undef FLT_ROUNDS
+#  undef FLT_RADIX
+#  undef FLT_MANT_DIG
+#  undef DBL_MANT_DIG
+#  undef LDBL_MANT_DIG
+#  undef DECIMAL_DIG
+#  undef FLT_DIG
+#  undef DBL_DIG
+#  undef LDBL_DIG
+#  undef FLT_MIN_EXP
+#  undef DBL_MIN_EXP
+#  undef LDBL_MIN_EXP
+#  undef FLT_MIN_10_EXP
+#  undef DBL_MIN_10_EXP
+#  undef LDBL_MIN_10_EXP
+#  undef FLT_MAX_EXP
+#  undef DBL_MAX_EXP
+#  undef LDBL_MAX_EXP
+#  undef FLT_MAX_10_EXP
+#  undef DBL_MAX_10_EXP
+#  undef LDBL_MAX_10_EXP
+#  undef FLT_MAX
+#  undef DBL_MAX
+#  undef LDBL_MAX
+#  undef FLT_EPSILON
+#  undef DBL_EPSILON
+#  undef LDBL_EPSILON
+#  undef FLT_MIN
+#  undef DBL_MIN
+#  undef LDBL_MIN
+#  if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#    undef FLT_TRUE_MIN
+#    undef DBL_TRUE_MIN
+#    undef LDBL_TRUE_MIN
+#  endif
+#endif
+
+/* Characteristics of floating point types, C99 5.2.4.2.2 */
+
+#define FLT_EVAL_METHOD __FLT_EVAL_METHOD__
+#define FLT_ROUNDS (__builtin_flt_rounds())
+#define FLT_RADIX __FLT_RADIX__
+
+#define FLT_MANT_DIG __FLT_MANT_DIG__
+#define DBL_MANT_DIG __DBL_MANT_DIG__
+#define LDBL_MANT_DIG __LDBL_MANT_DIG__
+
+#define DECIMAL_DIG __DECIMAL_DIG__
+
+#define FLT_DIG __FLT_DIG__
+#define DBL_DIG __DBL_DIG__
+#define LDBL_DIG __LDBL_DIG__
+
+#define FLT_MIN_EXP __FLT_MIN_EXP__
+#define DBL_MIN_EXP __DBL_MIN_EXP__
+#define LDBL_MIN_EXP __LDBL_MIN_EXP__
+
+#define FLT_MIN_10_EXP __FLT_MIN_10_EXP__
+#define DBL_MIN_10_EXP __DBL_MIN_10_EXP__
+#define LDBL_MIN_10_EXP __LDBL_MIN_10_EXP__
+
+#define FLT_MAX_EXP __FLT_MAX_EXP__
+#define DBL_MAX_EXP __DBL_MAX_EXP__
+#define LDBL_MAX_EXP __LDBL_MAX_EXP__
+
+#define FLT_MAX_10_EXP __FLT_MAX_10_EXP__
+#define DBL_MAX_10_EXP __DBL_MAX_10_EXP__
+#define LDBL_MAX_10_EXP __LDBL_MAX_10_EXP__
+
+#define FLT_MAX __FLT_MAX__
+#define DBL_MAX __DBL_MAX__
+#define LDBL_MAX __LDBL_MAX__
+
+#define FLT_EPSILON __FLT_EPSILON__
+#define DBL_EPSILON __DBL_EPSILON__
+#define LDBL_EPSILON __LDBL_EPSILON__
+
+#define FLT_MIN __FLT_MIN__
+#define DBL_MIN __DBL_MIN__
+#define LDBL_MIN __LDBL_MIN__
+
+#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#  define FLT_TRUE_MIN __FLT_DENORM_MIN__
+#  define DBL_TRUE_MIN __DBL_DENORM_MIN__
+#  define LDBL_TRUE_MIN __LDBL_DENORM_MIN__
+#endif
+
+#endif /* __FLOAT_H */
diff --git a/clang/lib/Headers/fma4intrin.h b/clang/lib/Headers/fma4intrin.h
new file mode 100644
index 0000000..c30920d
--- /dev/null
+++ b/clang/lib/Headers/fma4intrin.h
@@ -0,0 +1,231 @@
+/*===---- fma4intrin.h - FMA4 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <fma4intrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __FMA4INTRIN_H
+#define __FMA4INTRIN_H
+
+#ifndef __FMA4__
+# error "FMA4 instruction set is not enabled"
+#else
+
+#include <pmmintrin.h>
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_macc_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_macc_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_macc_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_macc_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_maddsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_maddsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msubadd_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msubadd_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubaddpd(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_macc_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_macc_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_msub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_msub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_nmacc_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_nmacc_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_nmsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_nmsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_maddsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_maddsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_msubadd_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_msubadd_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubaddpd256(__A, __B, __C);
+}
+
+#endif /* __FMA4__ */
+
+#endif /* __FMA4INTRIN_H */
diff --git a/clang/lib/Headers/immintrin.h b/clang/lib/Headers/immintrin.h
new file mode 100644
index 0000000..1605525
--- /dev/null
+++ b/clang/lib/Headers/immintrin.h
@@ -0,0 +1,75 @@
+/*===---- immintrin.h - Intel intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#define __IMMINTRIN_H
+
+#ifdef __MMX__
+#include <mmintrin.h>
+#endif
+
+#ifdef __SSE__
+#include <xmmintrin.h>
+#endif
+
+#ifdef __SSE2__
+#include <emmintrin.h>
+#endif
+
+#ifdef __SSE3__
+#include <pmmintrin.h>
+#endif
+
+#ifdef __SSSE3__
+#include <tmmintrin.h>
+#endif
+
+#if defined (__SSE4_2__) || defined (__SSE4_1__)
+#include <smmintrin.h>
+#endif
+
+#if defined (__AES__)
+#include <wmmintrin.h>
+#endif
+
+#ifdef __AVX__
+#include <avxintrin.h>
+#endif
+
+#ifdef __AVX2__
+#include <avx2intrin.h>
+#endif
+
+#ifdef __BMI__
+#include <bmiintrin.h>
+#endif
+
+#ifdef __BMI2__
+#include <bmi2intrin.h>
+#endif
+
+#ifdef __LZCNT__
+#include <lzcntintrin.h>
+#endif
+
+#endif /* __IMMINTRIN_H */
diff --git a/clang/lib/Headers/iso646.h b/clang/lib/Headers/iso646.h
new file mode 100644
index 0000000..dca13c5
--- /dev/null
+++ b/clang/lib/Headers/iso646.h
@@ -0,0 +1,43 @@
+/*===---- iso646.h - Standard header for alternate spellings of operators---===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ISO646_H
+#define __ISO646_H
+
+#ifndef __cplusplus
+#define and    &&
+#define and_eq &=
+#define bitand &
+#define bitor  |
+#define compl  ~
+#define not    !
+#define not_eq !=
+#define or     ||
+#define or_eq  |=
+#define xor    ^
+#define xor_eq ^=
+#endif
+
+#endif /* __ISO646_H */
diff --git a/clang/lib/Headers/limits.h b/clang/lib/Headers/limits.h
new file mode 100644
index 0000000..ecd09a4
--- /dev/null
+++ b/clang/lib/Headers/limits.h
@@ -0,0 +1,117 @@
+/*===---- limits.h - Standard header for integer sizes --------------------===*\
+ *
+ * Copyright (c) 2009 Chris Lattner
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __CLANG_LIMITS_H
+#define __CLANG_LIMITS_H
+
+/* The system's limits.h may, in turn, try to #include_next GCC's limits.h.
+   Avert this #include_next madness. */
+#if defined __GNUC__ && !defined _GCC_LIMITS_H_
+#define _GCC_LIMITS_H_
+#endif
+
+/* System headers include a number of constants from POSIX in <limits.h>.
+   Include it if we're hosted. */
+#if __STDC_HOSTED__ && \
+    defined(__has_include_next) && __has_include_next(<limits.h>)
+#include_next <limits.h>
+#endif
+
+/* Many system headers try to "help us out" by defining these.  No really, we
+   know how big each datatype is. */
+#undef  SCHAR_MIN
+#undef  SCHAR_MAX
+#undef  UCHAR_MAX
+#undef  SHRT_MIN
+#undef  SHRT_MAX
+#undef  USHRT_MAX
+#undef  INT_MIN
+#undef  INT_MAX
+#undef  UINT_MAX
+#undef  LONG_MIN
+#undef  LONG_MAX
+#undef  ULONG_MAX
+
+#undef  CHAR_BIT
+#undef  CHAR_MIN
+#undef  CHAR_MAX
+
+/* C90/99 5.2.4.2.1 */
+#define SCHAR_MAX __SCHAR_MAX__
+#define SHRT_MAX  __SHRT_MAX__
+#define INT_MAX   __INT_MAX__
+#define LONG_MAX  __LONG_MAX__
+
+#define SCHAR_MIN (-__SCHAR_MAX__-1)
+#define SHRT_MIN  (-__SHRT_MAX__ -1)
+#define INT_MIN   (-__INT_MAX__  -1)
+#define LONG_MIN  (-__LONG_MAX__ -1L)
+
+#define UCHAR_MAX (__SCHAR_MAX__*2  +1)
+#define USHRT_MAX (__SHRT_MAX__ *2  +1)
+#define UINT_MAX  (__INT_MAX__  *2U +1U)
+#define ULONG_MAX (__LONG_MAX__ *2UL+1UL)
+
+#ifndef MB_LEN_MAX
+#define MB_LEN_MAX 1
+#endif
+
+#define CHAR_BIT  __CHAR_BIT__
+
+#ifdef __CHAR_UNSIGNED__  /* -funsigned-char */
+#define CHAR_MIN 0
+#define CHAR_MAX UCHAR_MAX
+#else
+#define CHAR_MIN SCHAR_MIN
+#define CHAR_MAX __SCHAR_MAX__
+#endif
+
+/* C99 5.2.4.2.1: Added long long. */
+#if __STDC_VERSION__ >= 199901
+
+#undef  LLONG_MIN
+#undef  LLONG_MAX
+#undef  ULLONG_MAX
+
+#define LLONG_MAX  __LONG_LONG_MAX__
+#define LLONG_MIN  (-__LONG_LONG_MAX__-1LL)
+#define ULLONG_MAX (__LONG_LONG_MAX__*2ULL+1ULL)
+#endif
+
+/* LONG_LONG_MIN/LONG_LONG_MAX/ULONG_LONG_MAX are a GNU extension.  It's too bad
+   that we don't have something like #pragma poison that could be used to
+   deprecate a macro - the code should just use LLONG_MAX and friends.
+ */
+#if defined(__GNU_LIBRARY__) ? defined(__USE_GNU) : !defined(__STRICT_ANSI__)
+
+#undef   LONG_LONG_MIN
+#undef   LONG_LONG_MAX
+#undef   ULONG_LONG_MAX
+
+#define LONG_LONG_MAX  __LONG_LONG_MAX__
+#define LONG_LONG_MIN  (-__LONG_LONG_MAX__-1LL)
+#define ULONG_LONG_MAX (__LONG_LONG_MAX__*2ULL+1ULL)
+#endif
+
+#endif /* __CLANG_LIMITS_H */
diff --git a/clang/lib/Headers/lzcntintrin.h b/clang/lib/Headers/lzcntintrin.h
new file mode 100644
index 0000000..62ab5ca
--- /dev/null
+++ b/clang/lib/Headers/lzcntintrin.h
@@ -0,0 +1,55 @@
+/*===---- lzcntintrin.h - LZCNT intrinsics ---------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <lzcntintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __LZCNT__
+# error "LZCNT instruction is not enabled"
+#endif /* __LZCNT__ */
+
+#ifndef __LZCNTINTRIN_H
+#define __LZCNTINTRIN_H
+
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__lzcnt16(unsigned short __X)
+{
+  return __builtin_clzs(__X);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__lzcnt32(unsigned int __X)
+{
+  return __builtin_clz(__X);
+}
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__lzcnt64(unsigned long long __X)
+{
+  return __builtin_clzll(__X);
+}
+#endif
+
+#endif /* __LZCNTINTRIN_H */
diff --git a/clang/lib/Headers/mm3dnow.h b/clang/lib/Headers/mm3dnow.h
new file mode 100644
index 0000000..d5236f8
--- /dev/null
+++ b/clang/lib/Headers/mm3dnow.h
@@ -0,0 +1,161 @@
+/*===---- mm3dnow.h - 3DNow! intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _MM3DNOW_H_INCLUDED
+#define _MM3DNOW_H_INCLUDED
+
+#include <mmintrin.h>
+
+typedef float __v2sf __attribute__((__vector_size__(8)));
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_m_femms() {
+  __builtin_ia32_femms();
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pavgusb(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pavgusb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pf2id(__m64 __m) {
+  return (__m64)__builtin_ia32_pf2id((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfadd(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfadd((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpeq(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpeq((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpge(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpge((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpgt(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpgt((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmax(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmax((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmin(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmin((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmul(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmul((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcp(__m64 __m) {
+  return (__m64)__builtin_ia32_pfrcp((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcpit1(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrcpit1((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcpit2(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrcpit2((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrsqrt(__m64 __m) {
+  return (__m64)__builtin_ia32_pfrsqrt((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrsqrtit1(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrsqit1((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfsub(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfsub((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfsubr(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfsubr((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pi2fd(__m64 __m) {
+  return (__m64)__builtin_ia32_pi2fd((__v2si)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pmulhrw(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pmulhrw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pf2iw(__m64 __m) {
+  return (__m64)__builtin_ia32_pf2iw((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfnacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfnacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfpnacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfpnacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pi2fw(__m64 __m) {
+  return (__m64)__builtin_ia32_pi2fw((__v2si)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pswapdsf(__m64 __m) {
+  return (__m64)__builtin_ia32_pswapdsf((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pswapdsi(__m64 __m) {
+  return (__m64)__builtin_ia32_pswapdsi((__v2si)__m);
+}
+
+#endif
diff --git a/clang/lib/Headers/mm_malloc.h b/clang/lib/Headers/mm_malloc.h
new file mode 100644
index 0000000..5fa1761
--- /dev/null
+++ b/clang/lib/Headers/mm_malloc.h
@@ -0,0 +1,75 @@
+/*===---- mm_malloc.h - Allocating and Freeing Aligned Memory Blocks -------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __MM_MALLOC_H
+#define __MM_MALLOC_H
+
+#include <stdlib.h>
+
+#ifdef _WIN32
+#include <malloc.h>
+#else
+#ifndef __cplusplus
+extern int posix_memalign(void **memptr, size_t alignment, size_t size);
+#else
+// Some systems (e.g. those with GNU libc) declare posix_memalign with an
+// exception specifier. Via an "egregious workaround" in
+// Sema::CheckEquivalentExceptionSpec, Clang accepts the following as a valid
+// redeclaration of glibc's declaration.
+extern "C" int posix_memalign(void **memptr, size_t alignment, size_t size);
+#endif
+#endif
+
+#if !(defined(_WIN32) && defined(_mm_malloc))
+static __inline__ void *__attribute__((__always_inline__, __nodebug__,
+                                       __malloc__))
+_mm_malloc(size_t size, size_t align)
+{
+  if (align == 1) {
+    return malloc(size);
+  }
+
+  if (!(align & (align - 1)) && align < sizeof(void *))
+    align = sizeof(void *);
+
+  void *mallocedMemory;
+#if defined(__MINGW32__)
+  mallocedMemory = __mingw_aligned_malloc(size, align);
+#elif defined(_WIN32)
+  mallocedMemory = _aligned_malloc(size, align);
+#else
+  if (posix_memalign(&mallocedMemory, align, size))
+    return 0;
+#endif
+
+  return mallocedMemory;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_free(void *p)
+{
+  free(p);
+}
+#endif
+
+#endif /* __MM_MALLOC_H */
diff --git a/clang/lib/Headers/mmintrin.h b/clang/lib/Headers/mmintrin.h
new file mode 100644
index 0000000..986870a
--- /dev/null
+++ b/clang/lib/Headers/mmintrin.h
@@ -0,0 +1,503 @@
+/*===---- mmintrin.h - MMX intrinsics --------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __MMINTRIN_H
+#define __MMINTRIN_H
+
+#ifndef __MMX__
+#error "MMX instruction set not enabled"
+#else
+
+typedef long long __m64 __attribute__((__vector_size__(8)));
+
+typedef int __v2si __attribute__((__vector_size__(8)));
+typedef short __v4hi __attribute__((__vector_size__(8)));
+typedef char __v8qi __attribute__((__vector_size__(8)));
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_empty(void)
+{
+    __builtin_ia32_emms();
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_si64(int __i)
+{
+    return (__m64)__builtin_ia32_vec_init_v2si(__i, 0);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_si32(__m64 __m)
+{
+    return __builtin_ia32_vec_ext_v2si((__v2si)__m, 0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_m64(long long __i)
+{
+    return (__m64)__i;
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtm64_si64(__m64 __m)
+{
+    return (long long)__m;
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packsswb((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packssdw((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pu16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packuswb((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhbw((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhdq((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpcklbw((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpcklwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckldq((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pi8(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddsb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddsw((__v4hi)__m1, (__v4hi)__m2);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pu8(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddusb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pu16(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddusw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubw((__v4hi)__m1, (__v4hi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubsb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubsw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pu8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubusb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pu16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubusw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_madd_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pmaddwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pmulhw((__v4hi)__m1, (__v4hi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_pi16(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_pmullw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psllw((__v4hi)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psllwi((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_pslld((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_pslldi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_si64(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psllq(__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_si64(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psllqi(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sra_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psraw((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srai_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrawi((__v4hi)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sra_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrad((__v2si)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srai_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psradi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrlw((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrlwi((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrld((__v2si)__m, __count);       
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrldi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_si64(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrlq(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_si64(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrlqi(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_and_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pand(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pandn(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_or_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_por(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_xor_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pxor(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_si64(void)
+{
+    return (__m64){ 0LL };
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi32(int __i1, int __i0)
+{
+    return (__m64)__builtin_ia32_vec_init_v2si(__i0, __i1);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi16(short __s3, short __s2, short __s1, short __s0)
+{
+    return (__m64)__builtin_ia32_vec_init_v4hi(__s0, __s1, __s2, __s3);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi8(char __b7, char __b6, char __b5, char __b4, char __b3, char __b2,
+            char __b1, char __b0)
+{
+    return (__m64)__builtin_ia32_vec_init_v8qi(__b0, __b1, __b2, __b3,
+                                               __b4, __b5, __b6, __b7);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi32(int __i)
+{
+    return _mm_set_pi32(__i, __i);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi16(short __w)
+{
+    return _mm_set_pi16(__w, __w, __w, __w);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi8(char __b)
+{
+    return _mm_set_pi8(__b, __b, __b, __b, __b, __b, __b, __b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi32(int __i0, int __i1)
+{
+    return _mm_set_pi32(__i1, __i0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi16(short __w0, short __w1, short __w2, short __w3)
+{
+    return _mm_set_pi16(__w3, __w2, __w1, __w0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi8(char __b0, char __b1, char __b2, char __b3, char __b4, char __b5,
+             char __b6, char __b7)
+{
+    return _mm_set_pi8(__b7, __b6, __b5, __b4, __b3, __b2, __b1, __b0);
+}
+
+
+/* Aliases for compatibility. */
+#define _m_empty _mm_empty
+#define _m_from_int _mm_cvtsi32_si64
+#define _m_to_int _mm_cvtsi64_si32
+#define _m_packsswb _mm_packs_pi16
+#define _m_packssdw _mm_packs_pi32
+#define _m_packuswb _mm_packs_pu16
+#define _m_punpckhbw _mm_unpackhi_pi8
+#define _m_punpckhwd _mm_unpackhi_pi16
+#define _m_punpckhdq _mm_unpackhi_pi32
+#define _m_punpcklbw _mm_unpacklo_pi8
+#define _m_punpcklwd _mm_unpacklo_pi16
+#define _m_punpckldq _mm_unpacklo_pi32
+#define _m_paddb _mm_add_pi8
+#define _m_paddw _mm_add_pi16
+#define _m_paddd _mm_add_pi32
+#define _m_paddsb _mm_adds_pi8
+#define _m_paddsw _mm_adds_pi16
+#define _m_paddusb _mm_adds_pu8
+#define _m_paddusw _mm_adds_pu16
+#define _m_psubb _mm_sub_pi8
+#define _m_psubw _mm_sub_pi16
+#define _m_psubd _mm_sub_pi32
+#define _m_psubsb _mm_subs_pi8
+#define _m_psubsw _mm_subs_pi16
+#define _m_psubusb _mm_subs_pu8
+#define _m_psubusw _mm_subs_pu16
+#define _m_pmaddwd _mm_madd_pi16
+#define _m_pmulhw _mm_mulhi_pi16
+#define _m_pmullw _mm_mullo_pi16
+#define _m_psllw _mm_sll_pi16
+#define _m_psllwi _mm_slli_pi16
+#define _m_pslld _mm_sll_pi32
+#define _m_pslldi _mm_slli_pi32
+#define _m_psllq _mm_sll_si64
+#define _m_psllqi _mm_slli_si64
+#define _m_psraw _mm_sra_pi16
+#define _m_psrawi _mm_srai_pi16
+#define _m_psrad _mm_sra_pi32
+#define _m_psradi _mm_srai_pi32
+#define _m_psrlw _mm_srl_pi16
+#define _m_psrlwi _mm_srli_pi16
+#define _m_psrld _mm_srl_pi32
+#define _m_psrldi _mm_srli_pi32
+#define _m_psrlq _mm_srl_si64
+#define _m_psrlqi _mm_srli_si64
+#define _m_pand _mm_and_si64
+#define _m_pandn _mm_andnot_si64
+#define _m_por _mm_or_si64
+#define _m_pxor _mm_xor_si64
+#define _m_pcmpeqb _mm_cmpeq_pi8
+#define _m_pcmpeqw _mm_cmpeq_pi16
+#define _m_pcmpeqd _mm_cmpeq_pi32
+#define _m_pcmpgtb _mm_cmpgt_pi8
+#define _m_pcmpgtw _mm_cmpgt_pi16
+#define _m_pcmpgtd _mm_cmpgt_pi32
+
+#endif /* __MMX__ */
+
+#endif /* __MMINTRIN_H */
+
diff --git a/clang/lib/Headers/module.map b/clang/lib/Headers/module.map
new file mode 100644
index 0000000..418ba50
--- /dev/null
+++ b/clang/lib/Headers/module.map
@@ -0,0 +1,108 @@
+module _Builtin_intrinsics [system] {
+  explicit module altivec {
+    requires altivec
+    header "altivec.h"
+  }
+
+  explicit module intel {
+    requires x86
+    export *
+
+    header "immintrin.h"
+    header "x86intrin.h"
+
+    explicit module mm_malloc {
+      header "mm_malloc.h"
+      export * // note: for <stdlib.h> dependency
+    }
+
+    explicit module cpuid {
+      header "cpuid.h"
+    }
+
+    explicit module mmx {
+      requires mmx
+      header "mmintrin.h"
+    }
+
+    explicit module sse {
+      requires sse
+      export mmx
+      export * // note: for hackish <emmintrin.h> dependency
+      header "xmmintrin.h"
+    }
+
+    explicit module sse2 {
+      requires sse2
+      export sse
+      header "emmintrin.h"
+    }
+
+    explicit module sse3 {
+      requires sse3
+      export sse2
+      header "pmmintrin.h"
+    }
+
+    explicit module ssse3 {
+      requires ssse3
+      export sse3
+      header "tmmintrin.h"
+    }
+
+    explicit module sse4_1 {
+      requires sse41
+      export ssse3
+      header "smmintrin.h"
+    }
+
+    explicit module sse4_2 {
+      requires sse42
+      export sse4_1
+      header "nmmintrin.h"
+    }
+
+    explicit module avx {
+      requires avx
+      export sse4_2
+      header "avxintrin.h"
+    }
+
+    explicit module avx2 {
+      requires avx2
+      export avx
+      header "avx2intrin.h"
+    }
+
+    explicit module bmi {
+      requires bmi
+      header "bmiintrin.h"
+    }
+
+    explicit module bmi2 {
+      requires bmi2
+      header "bmi2intrin.h"
+    }
+
+    explicit module fma4 {
+      requires fma4
+      export sse3
+      header "fma4intrin.h"
+    }
+
+    explicit module lzcnt {
+      requires lzcnt
+      header "lzcntintrin.h"
+    }
+
+    explicit module popcnt {
+      requires popcnt
+      header "popcntintrin.h"
+    }
+
+    explicit module mm3dnow {
+      requires mm3dnow
+      header "mm3dnow.h"
+    }
+  }
+}
diff --git a/clang/lib/Headers/nmmintrin.h b/clang/lib/Headers/nmmintrin.h
new file mode 100644
index 0000000..f12622d
--- /dev/null
+++ b/clang/lib/Headers/nmmintrin.h
@@ -0,0 +1,35 @@
+/*===---- nmmintrin.h - SSE4 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _NMMINTRIN_H
+#define _NMMINTRIN_H
+
+#ifndef __SSE4_2__
+#error "SSE4.2 instruction set not enabled"
+#else
+
+/* To match expectations of gcc we put the sse4.2 definitions into smmintrin.h,
+   just include it now then.  */
+#include <smmintrin.h>
+#endif /* __SSE4_2__ */
+#endif /* _NMMINTRIN_H */
diff --git a/clang/lib/Headers/pmmintrin.h b/clang/lib/Headers/pmmintrin.h
new file mode 100644
index 0000000..5f9b097
--- /dev/null
+++ b/clang/lib/Headers/pmmintrin.h
@@ -0,0 +1,117 @@
+/*===---- pmmintrin.h - SSE3 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __PMMINTRIN_H
+#define __PMMINTRIN_H
+
+#ifndef __SSE3__
+#error "SSE3 instruction set not enabled"
+#else
+
+#include <emmintrin.h>
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_lddqu_si128(__m128i const *p)
+{
+  return (__m128i)__builtin_ia32_lddqu((char const *)p);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_addsub_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_addsubps(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_haddps(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_hsubps(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movehdup_ps(__m128 a)
+{
+  return __builtin_shufflevector(a, a, 1, 1, 3, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_moveldup_ps(__m128 a)
+{
+  return __builtin_shufflevector(a, a, 0, 0, 2, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_addsub_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_addsubpd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_haddpd(a, b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pd(__m128d a, __m128d b)
+{
+  return __builtin_ia32_hsubpd(a, b);
+}
+
+#define        _mm_loaddup_pd(dp)        _mm_load1_pd(dp)
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_movedup_pd(__m128d a)
+{
+  return __builtin_shufflevector(a, a, 0, 0);
+}
+
+#define _MM_DENORMALS_ZERO_ON   (0x0040)
+#define _MM_DENORMALS_ZERO_OFF  (0x0000)
+
+#define _MM_DENORMALS_ZERO_MASK (0x0040)
+
+#define _MM_GET_DENORMALS_ZERO_MODE() (_mm_getcsr() & _MM_DENORMALS_ZERO_MASK)
+#define _MM_SET_DENORMALS_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_DENORMALS_ZERO_MASK) | (x)))
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_monitor(void const *p, unsigned extensions, unsigned hints)
+{
+  __builtin_ia32_monitor((void *)p, extensions, hints);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_mwait(unsigned extensions, unsigned hints)
+{
+  __builtin_ia32_mwait(extensions, hints);
+}
+
+#endif /* __SSE3__ */
+
+#endif /* __PMMINTRIN_H */
diff --git a/clang/lib/Headers/popcntintrin.h b/clang/lib/Headers/popcntintrin.h
new file mode 100644
index 0000000..d439daa
--- /dev/null
+++ b/clang/lib/Headers/popcntintrin.h
@@ -0,0 +1,45 @@
+/*===---- popcntintrin.h - POPCNT intrinsics -------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __POPCNT__
+#error "POPCNT instruction set not enabled"
+#endif
+
+#ifndef _POPCNTINTRIN_H
+#define _POPCNTINTRIN_H
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_popcnt_u32(unsigned int __A)
+{
+  return __builtin_popcount(__A);
+}
+
+#ifdef __x86_64__
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_popcnt_u64(unsigned long long __A)
+{
+  return __builtin_popcountll(__A);
+}
+#endif /* __x86_64__ */
+
+#endif /* _POPCNTINTRIN_H */
diff --git a/clang/lib/Headers/smmintrin.h b/clang/lib/Headers/smmintrin.h
new file mode 100644
index 0000000..2fab50e
--- /dev/null
+++ b/clang/lib/Headers/smmintrin.h
@@ -0,0 +1,467 @@
+/*===---- smmintrin.h - SSE4 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _SMMINTRIN_H
+#define _SMMINTRIN_H
+
+#ifndef __SSE4_1__
+#error "SSE4.1 instruction set not enabled"
+#else
+
+#include <tmmintrin.h>
+
+/* SSE4 Rounding macros. */
+#define _MM_FROUND_TO_NEAREST_INT    0x00
+#define _MM_FROUND_TO_NEG_INF        0x01
+#define _MM_FROUND_TO_POS_INF        0x02
+#define _MM_FROUND_TO_ZERO           0x03
+#define _MM_FROUND_CUR_DIRECTION     0x04
+
+#define _MM_FROUND_RAISE_EXC         0x00
+#define _MM_FROUND_NO_EXC            0x08
+
+#define _MM_FROUND_NINT      (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEAREST_INT)
+#define _MM_FROUND_FLOOR     (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEG_INF)
+#define _MM_FROUND_CEIL      (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_POS_INF)
+#define _MM_FROUND_TRUNC     (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_ZERO)
+#define _MM_FROUND_RINT      (_MM_FROUND_RAISE_EXC | _MM_FROUND_CUR_DIRECTION)
+#define _MM_FROUND_NEARBYINT (_MM_FROUND_NO_EXC | _MM_FROUND_CUR_DIRECTION)
+
+#define _mm_ceil_ps(X)       _mm_round_ps((X), _MM_FROUND_CEIL)
+#define _mm_ceil_pd(X)       _mm_round_pd((X), _MM_FROUND_CEIL)
+#define _mm_ceil_ss(X, Y)    _mm_round_ss((X), (Y), _MM_FROUND_CEIL)
+#define _mm_ceil_sd(X, Y)    _mm_round_sd((X), (Y), _MM_FROUND_CEIL)
+
+#define _mm_floor_ps(X)      _mm_round_ps((X), _MM_FROUND_FLOOR)
+#define _mm_floor_pd(X)      _mm_round_pd((X), _MM_FROUND_FLOOR)
+#define _mm_floor_ss(X, Y)   _mm_round_ss((X), (Y), _MM_FROUND_FLOOR)
+#define _mm_floor_sd(X, Y)   _mm_round_sd((X), (Y), _MM_FROUND_FLOOR)
+
+#define _mm_round_ps(X, M) __extension__ ({ \
+  __m128 __X = (X); \
+  (__m128) __builtin_ia32_roundps((__v4sf)__X, (M)); })
+
+#define _mm_round_ss(X, Y, M) __extension__ ({ \
+  __m128 __X = (X); \
+  __m128 __Y = (Y); \
+  (__m128) __builtin_ia32_roundss((__v4sf)__X, (__v4sf)__Y, (M)); })
+
+#define _mm_round_pd(X, M) __extension__ ({ \
+  __m128d __X = (X); \
+  (__m128d) __builtin_ia32_roundpd((__v2df)__X, (M)); })
+
+#define _mm_round_sd(X, Y, M) __extension__ ({ \
+  __m128d __X = (X); \
+  __m128d __Y = (Y); \
+  (__m128d) __builtin_ia32_roundsd((__v2df)__X, (__v2df)__Y, (M)); })
+
+/* SSE4 Packed Blending Intrinsics.  */
+#define _mm_blend_pd(V1, V2, M) __extension__ ({ \
+  __m128d __V1 = (V1); \
+  __m128d __V2 = (V2); \
+  (__m128d) __builtin_ia32_blendpd ((__v2df)__V1, (__v2df)__V2, (M)); })
+
+#define _mm_blend_ps(V1, V2, M) __extension__ ({ \
+  __m128 __V1 = (V1); \
+  __m128 __V2 = (V2); \
+  (__m128) __builtin_ia32_blendps ((__v4sf)__V1, (__v4sf)__V2, (M)); })
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_pd (__m128d __V1, __m128d __V2, __m128d __M)
+{
+  return (__m128d) __builtin_ia32_blendvpd ((__v2df)__V1, (__v2df)__V2,
+                                            (__v2df)__M);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_ps (__m128 __V1, __m128 __V2, __m128 __M)
+{
+  return (__m128) __builtin_ia32_blendvps ((__v4sf)__V1, (__v4sf)__V2,
+                                           (__v4sf)__M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_epi8 (__m128i __V1, __m128i __V2, __m128i __M)
+{
+  return (__m128i) __builtin_ia32_pblendvb128 ((__v16qi)__V1, (__v16qi)__V2,
+                                               (__v16qi)__M);
+}
+
+#define _mm_blend_epi16(V1, V2, M) __extension__ ({ \
+  __m128i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m128i) __builtin_ia32_pblendw128 ((__v8hi)__V1, (__v8hi)__V2, (M)); })
+
+/* SSE4 Dword Multiply Instructions.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) ((__v4si)__V1 * (__v4si)__V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mul_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmuldq128 ((__v4si)__V1, (__v4si)__V2);
+}
+
+/* SSE4 Floating Point Dot Product Instructions.  */
+#define _mm_dp_ps(X, Y, M) __extension__ ({ \
+  __m128 __X = (X); \
+  __m128 __Y = (Y); \
+  (__m128) __builtin_ia32_dpps((__v4sf)__X, (__v4sf)__Y, (M)); })
+
+#define _mm_dp_pd(X, Y, M) __extension__ ({\
+  __m128d __X = (X); \
+  __m128d __Y = (Y); \
+  (__m128d) __builtin_ia32_dppd((__v2df)__X, (__v2df)__Y, (M)); })
+
+/* SSE4 Streaming Load Hint Instruction.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_stream_load_si128 (__m128i *__V)
+{
+  return (__m128i) __builtin_ia32_movntdqa ((__v2di *) __V);
+}
+
+/* SSE4 Packed Integer Min/Max Instructions.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi8 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminsb128 ((__v16qi) __V1, (__v16qi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi8 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxsb128 ((__v16qi) __V1, (__v16qi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu16 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminuw128 ((__v8hi) __V1, (__v8hi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu16 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxuw128 ((__v8hi) __V1, (__v8hi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminsd128 ((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxsd128 ((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminud128((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxud128((__v4si) __V1, (__v4si) __V2);
+}
+
+/* SSE4 Insertion and Extraction from XMM Register Instructions.  */
+#define _mm_insert_ps(X, Y, N) __builtin_ia32_insertps128((X), (Y), (N))
+#define _mm_extract_ps(X, N) (__extension__                      \
+                              ({ union { int i; float f; } __t;  \
+                                 __v4sf __a = (__v4sf)(X);       \
+                                 __t.f = __a[N];                 \
+                                 __t.i;}))
+
+/* Miscellaneous insert and extract macros.  */
+/* Extract a single-precision float from X at index N into D.  */
+#define _MM_EXTRACT_FLOAT(D, X, N) (__extension__ ({ __v4sf __a = (__v4sf)(X); \
+                                                    (D) = __a[N]; }))
+                                                    
+/* Or together 2 sets of indexes (X and Y) with the zeroing bits (Z) to create
+   an index suitable for _mm_insert_ps.  */
+#define _MM_MK_INSERTPS_NDX(X, Y, Z) (((X) << 6) | ((Y) << 4) | (Z))
+                                           
+/* Extract a float from X at index N into the first index of the return.  */
+#define _MM_PICK_OUT_PS(X, N) _mm_insert_ps (_mm_setzero_ps(), (X),   \
+                                             _MM_MK_INSERTPS_NDX((N), 0, 0x0e))
+                                             
+/* Insert int into packed integer array at index.  */
+#define _mm_insert_epi8(X, I, N) (__extension__ ({ __v16qi __a = (__v16qi)(X); \
+                                                   __a[(N)] = (I);             \
+                                                   __a;}))
+#define _mm_insert_epi32(X, I, N) (__extension__ ({ __v4si __a = (__v4si)(X); \
+                                                    __a[(N)] = (I);           \
+                                                    __a;}))
+#ifdef __x86_64__
+#define _mm_insert_epi64(X, I, N) (__extension__ ({ __v2di __a = (__v2di)(X); \
+                                                    __a[(N)] = (I);           \
+                                                    __a;}))
+#endif /* __x86_64__ */
+
+/* Extract int from packed integer array at index.  This returns the element
+ * as a zero extended value, so it is unsigned.
+ */
+#define _mm_extract_epi8(X, N) (__extension__ ({ __v16qi __a = (__v16qi)(X); \
+                                                 (unsigned char)__a[(N)];}))
+#define _mm_extract_epi32(X, N) (__extension__ ({ __v4si __a = (__v4si)(X); \
+                                                  (unsigned)__a[(N)];}))
+#ifdef __x86_64__
+#define _mm_extract_epi64(X, N) (__extension__ ({ __v2di __a = (__v2di)(X); \
+                                                  __a[(N)];}))
+#endif /* __x86_64 */
+
+/* SSE4 128-bit Packed Integer Comparisons.  */
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestz128((__v2di)__M, (__v2di)__V);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestc128((__v2di)__M, (__v2di)__V);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestnzc128((__v2di)__M, (__v2di)__V);
+}
+
+#define _mm_test_all_ones(V) _mm_testc_si128((V), _mm_cmpeq_epi32((V), (V)))
+#define _mm_test_mix_ones_zeros(M, V) _mm_testnzc_si128((M), (V))
+#define _mm_test_all_zeros(M, V) _mm_testz_si128 ((M), (V))
+
+/* SSE4 64-bit Packed Integer Comparisons.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi64(__m128i __V1, __m128i __V2)
+{
+  return (__m128i)((__v2di)__V1 == (__v2di)__V2);
+}
+
+/* SSE4 Packed Integer Sign-Extension.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbw128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbd128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbq128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi16_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxwd128((__v8hi) __V); 
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi16_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxwq128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxdq128((__v4si)__V);
+}
+
+/* SSE4 Packed Integer Zero-Extension.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbw128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbd128((__v16qi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbq128((__v16qi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu16_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxwd128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu16_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxwq128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu32_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxdq128((__v4si)__V);
+}
+
+/* SSE4 Pack with Unsigned Saturation.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packus_epi32(__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_packusdw128((__v4si)__V1, (__v4si)__V2);
+}
+
+/* SSE4 Multiple Packed Sums of Absolute Difference.  */
+#define _mm_mpsadbw_epu8(X, Y, M) __extension__ ({ \
+  __m128i __X = (X); \
+  __m128i __Y = (Y); \
+  (__m128i) __builtin_ia32_mpsadbw128((__v16qi)__X, (__v16qi)__Y, (M)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_minpos_epu16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_phminposuw128((__v8hi)__V);
+}
+
+/* These definitions are normally in nmmintrin.h, but gcc puts them in here
+   so we'll do the same.  */
+#ifdef __SSE4_2__
+
+/* These specify the type of data that we're comparing.  */
+#define _SIDD_UBYTE_OPS                 0x00
+#define _SIDD_UWORD_OPS                 0x01
+#define _SIDD_SBYTE_OPS                 0x02
+#define _SIDD_SWORD_OPS                 0x03
+
+/* These specify the type of comparison operation.  */
+#define _SIDD_CMP_EQUAL_ANY             0x00
+#define _SIDD_CMP_RANGES                0x04
+#define _SIDD_CMP_EQUAL_EACH            0x08
+#define _SIDD_CMP_EQUAL_ORDERED         0x0c
+
+/* These macros specify the polarity of the operation.  */
+#define _SIDD_POSITIVE_POLARITY         0x00
+#define _SIDD_NEGATIVE_POLARITY         0x10
+#define _SIDD_MASKED_POSITIVE_POLARITY  0x20
+#define _SIDD_MASKED_NEGATIVE_POLARITY  0x30
+
+/* These macros are used in _mm_cmpXstri() to specify the return.  */
+#define _SIDD_LEAST_SIGNIFICANT         0x00
+#define _SIDD_MOST_SIGNIFICANT          0x40
+
+/* These macros are used in _mm_cmpXstri() to specify the return.  */
+#define _SIDD_BIT_MASK                  0x00
+#define _SIDD_UNIT_MASK                 0x40
+
+/* SSE4.2 Packed Comparison Intrinsics.  */
+#define _mm_cmpistrm(A, B, M) __builtin_ia32_pcmpistrm128((A), (B), (M))
+#define _mm_cmpistri(A, B, M) __builtin_ia32_pcmpistri128((A), (B), (M))
+
+#define _mm_cmpestrm(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestrm128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestri(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestri128((A), (LA), (B), (LB), (M))
+     
+/* SSE4.2 Packed Comparison Intrinsics and EFlag Reading.  */
+#define _mm_cmpistra(A, B, M) \
+     __builtin_ia32_pcmpistria128((A), (B), (M))
+#define _mm_cmpistrc(A, B, M) \
+     __builtin_ia32_pcmpistric128((A), (B), (M))
+#define _mm_cmpistro(A, B, M) \
+     __builtin_ia32_pcmpistrio128((A), (B), (M))
+#define _mm_cmpistrs(A, B, M) \
+     __builtin_ia32_pcmpistris128((A), (B), (M))
+#define _mm_cmpistrz(A, B, M) \
+     __builtin_ia32_pcmpistriz128((A), (B), (M))
+
+#define _mm_cmpestra(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestria128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrc(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestric128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestro(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestrio128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrs(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestris128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrz(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestriz128((A), (LA), (B), (LB), (M))
+
+/* SSE4.2 Compare Packed Data -- Greater Than.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi64(__m128i __V1, __m128i __V2)
+{
+  return (__m128i)((__v2di)__V1 > (__v2di)__V2);
+}
+
+/* SSE4.2 Accumulate CRC32.  */
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u8(unsigned int __C, unsigned char __D)
+{
+  return __builtin_ia32_crc32qi(__C, __D);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u16(unsigned int __C, unsigned short __D)
+{
+  return __builtin_ia32_crc32hi(__C, __D);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u32(unsigned int __C, unsigned int __D)
+{
+  return __builtin_ia32_crc32si(__C, __D);
+}
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u64(unsigned long long __C, unsigned long long __D)
+{
+  return __builtin_ia32_crc32di(__C, __D);
+}
+#endif /* __x86_64__ */
+
+#ifdef __POPCNT__
+#include <popcntintrin.h>
+#endif
+
+#endif /* __SSE4_2__ */
+#endif /* __SSE4_1__ */
+
+#endif /* _SMMINTRIN_H */
diff --git a/clang/lib/Headers/stdalign.h b/clang/lib/Headers/stdalign.h
new file mode 100644
index 0000000..e7fbfa0
--- /dev/null
+++ b/clang/lib/Headers/stdalign.h
@@ -0,0 +1,30 @@
+/*===---- stdalign.h - Standard header for alignment ------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDALIGN_H
+#define __STDALIGN_H
+
+#define alignas _Alignas
+#define __alignas_is_defined 1
+
+#endif /* __STDALIGN_H */
diff --git a/clang/lib/Headers/stdarg.h b/clang/lib/Headers/stdarg.h
new file mode 100644
index 0000000..2957bf0
--- /dev/null
+++ b/clang/lib/Headers/stdarg.h
@@ -0,0 +1,50 @@
+/*===---- stdarg.h - Variable argument handling ----------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDARG_H
+#define __STDARG_H
+
+#ifndef _VA_LIST
+typedef __builtin_va_list va_list;
+#define _VA_LIST
+#endif
+#define va_start(ap, param) __builtin_va_start(ap, param)
+#define va_end(ap)          __builtin_va_end(ap)
+#define va_arg(ap, type)    __builtin_va_arg(ap, type)
+
+/* GCC always defines __va_copy, but does not define va_copy unless in c99 mode
+ * or -ansi is not specified, since it was not part of C90.
+ */
+#define __va_copy(d,s) __builtin_va_copy(d,s)
+
+#if __STDC_VERSION__ >= 199900L || __cplusplus >= 201103L || !defined(__STRICT_ANSI__)
+#define va_copy(dest, src)  __builtin_va_copy(dest, src)
+#endif
+
+/* Hack required to make standard headers work, at least on Ubuntu */
+#define __GNUC_VA_LIST 1
+typedef __builtin_va_list __gnuc_va_list;
+
+#endif /* __STDARG_H */
diff --git a/clang/lib/Headers/stdbool.h b/clang/lib/Headers/stdbool.h
new file mode 100644
index 0000000..0467893
--- /dev/null
+++ b/clang/lib/Headers/stdbool.h
@@ -0,0 +1,44 @@
+/*===---- stdbool.h - Standard header for booleans -------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDBOOL_H
+#define __STDBOOL_H
+
+/* Don't define bool, true, and false in C++, except as a GNU extension. */
+#ifndef __cplusplus
+#define bool _Bool
+#define true 1
+#define false 0
+#elif defined(__GNUC__) && !defined(__STRICT_ANSI__)
+/* Define _Bool, bool, false, true as a GNU extension. */
+#define _Bool bool
+#define bool  bool
+#define false false
+#define true  true
+#endif
+
+#define __bool_true_false_are_defined 1
+
+#endif /* __STDBOOL_H */
diff --git a/clang/lib/Headers/stddef.h b/clang/lib/Headers/stddef.h
new file mode 100644
index 0000000..9e87ee8
--- /dev/null
+++ b/clang/lib/Headers/stddef.h
@@ -0,0 +1,64 @@
+/*===---- stddef.h - Basic type definitions --------------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDDEF_H
+#define __STDDEF_H
+
+#ifndef _PTRDIFF_T
+#define _PTRDIFF_T
+typedef __typeof__(((int*)0)-((int*)0)) ptrdiff_t;
+#endif
+#ifndef _SIZE_T
+#define _SIZE_T
+typedef __typeof__(sizeof(int)) size_t;
+#endif
+#ifndef __cplusplus
+#ifndef _WCHAR_T
+#define _WCHAR_T
+typedef __WCHAR_TYPE__ wchar_t;
+#endif
+#endif
+
+#undef NULL
+#ifdef __cplusplus
+#undef __null  // VC++ hack.
+#define NULL __null
+#else
+#define NULL ((void*)0)
+#endif
+
+#define offsetof(t, d) __builtin_offsetof(t, d)
+
+#endif /* __STDDEF_H */
+
+/* Some C libraries expect to see a wint_t here. Others (notably MinGW) will use
+__WINT_TYPE__ directly; accommodate both by requiring __need_wint_t */
+#if defined(__need_wint_t)
+#if !defined(_WINT_T)
+#define _WINT_T
+typedef __WINT_TYPE__ wint_t;
+#endif /* _WINT_T */
+#undef __need_wint_t
+#endif /* __need_wint_t */
diff --git a/clang/lib/Headers/stdint.h b/clang/lib/Headers/stdint.h
new file mode 100644
index 0000000..6f1a876
--- /dev/null
+++ b/clang/lib/Headers/stdint.h
@@ -0,0 +1,661 @@
+/*===---- stdint.h - Standard header for sized integer types --------------===*\
+ *
+ * Copyright (c) 2009 Chris Lattner
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __CLANG_STDINT_H
+#define __CLANG_STDINT_H
+
+/* If we're hosted, fall back to the system's stdint.h, which might have
+ * additional definitions.
+ */
+#if __STDC_HOSTED__ && \
+    defined(__has_include_next) && __has_include_next(<stdint.h>)
+# include_next <stdint.h>
+#else
+
+/* C99 7.18.1.1 Exact-width integer types.
+ * C99 7.18.1.2 Minimum-width integer types.
+ * C99 7.18.1.3 Fastest minimum-width integer types.
+ *
+ * The standard requires that exact-width type be defined for 8-, 16-, 32-, and 
+ * 64-bit types if they are implemented. Other exact width types are optional.
+ * This implementation defines an exact-width types for every integer width
+ * that is represented in the standard integer types.
+ *
+ * The standard also requires minimum-width types be defined for 8-, 16-, 32-,
+ * and 64-bit widths regardless of whether there are corresponding exact-width
+ * types. 
+ *
+ * To accommodate targets that are missing types that are exactly 8, 16, 32, or
+ * 64 bits wide, this implementation takes an approach of cascading
+ * redefintions, redefining __int_leastN_t to successively smaller exact-width
+ * types. It is therefore important that the types are defined in order of
+ * descending widths.
+ *
+ * We currently assume that the minimum-width types and the fastest
+ * minimum-width types are the same. This is allowed by the standard, but is
+ * suboptimal.
+ *
+ * In violation of the standard, some targets do not implement a type that is
+ * wide enough to represent all of the required widths (8-, 16-, 32-, 64-bit).  
+ * To accommodate these targets, a required minimum-width type is only
+ * defined if there exists an exact-width type of equal or greater width.
+ */
+
+#ifdef __INT64_TYPE__
+# ifndef __int8_t_defined /* glibc sys/types.h also defines int64_t*/
+typedef signed __INT64_TYPE__ int64_t;
+# endif /* __int8_t_defined */
+typedef unsigned __INT64_TYPE__ uint64_t;
+# define __int_least64_t int64_t
+# define __uint_least64_t uint64_t
+# define __int_least32_t int64_t
+# define __uint_least32_t uint64_t
+# define __int_least16_t int64_t
+# define __uint_least16_t uint64_t
+# define __int_least8_t int64_t
+# define __uint_least8_t uint64_t
+#endif /* __INT64_TYPE__ */
+
+#ifdef __int_least64_t
+typedef __int_least64_t int_least64_t;
+typedef __uint_least64_t uint_least64_t;
+typedef __int_least64_t int_fast64_t;
+typedef __uint_least64_t uint_fast64_t;
+#endif /* __int_least64_t */
+
+#ifdef __INT56_TYPE__
+typedef signed __INT56_TYPE__ int56_t;
+typedef unsigned __INT56_TYPE__ uint56_t;
+typedef int56_t int_least56_t;
+typedef uint56_t uint_least56_t;
+typedef int56_t int_fast56_t;
+typedef uint56_t uint_fast56_t;
+# define __int_least32_t int56_t
+# define __uint_least32_t uint56_t
+# define __int_least16_t int56_t
+# define __uint_least16_t uint56_t
+# define __int_least8_t int56_t
+# define __uint_least8_t uint56_t
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+typedef signed __INT48_TYPE__ int48_t;
+typedef unsigned __INT48_TYPE__ uint48_t;
+typedef int48_t int_least48_t;
+typedef uint48_t uint_least48_t;
+typedef int48_t int_fast48_t;
+typedef uint48_t uint_fast48_t;
+# define __int_least32_t int48_t
+# define __uint_least32_t uint48_t
+# define __int_least16_t int48_t
+# define __uint_least16_t uint48_t
+# define __int_least8_t int48_t
+# define __uint_least8_t uint48_t
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+typedef signed __INT40_TYPE__ int40_t;
+typedef unsigned __INT40_TYPE__ uint40_t;
+typedef int40_t int_least40_t;
+typedef uint40_t uint_least40_t;
+typedef int40_t int_fast40_t;
+typedef uint40_t uint_fast40_t;
+# define __int_least32_t int40_t
+# define __uint_least32_t uint40_t
+# define __int_least16_t int40_t
+# define __uint_least16_t uint40_t
+# define __int_least8_t int40_t
+# define __uint_least8_t uint40_t
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+
+# ifndef __int8_t_defined /* glibc sys/types.h also defines int32_t*/
+typedef signed __INT32_TYPE__ int32_t;
+# endif /* __int8_t_defined */
+
+# ifndef __uint32_t_defined  /* more glibc compatibility */
+# define __uint32_t_defined
+typedef unsigned __INT32_TYPE__ uint32_t;
+# endif /* __uint32_t_defined */
+
+# define __int_least32_t int32_t
+# define __uint_least32_t uint32_t
+# define __int_least16_t int32_t
+# define __uint_least16_t uint32_t
+# define __int_least8_t int32_t
+# define __uint_least8_t uint32_t
+#endif /* __INT32_TYPE__ */
+
+#ifdef __int_least32_t
+typedef __int_least32_t int_least32_t;
+typedef __uint_least32_t uint_least32_t;
+typedef __int_least32_t int_fast32_t;
+typedef __uint_least32_t uint_fast32_t;
+#endif /* __int_least32_t */
+
+#ifdef __INT24_TYPE__
+typedef signed __INT24_TYPE__ int24_t;
+typedef unsigned __INT24_TYPE__ uint24_t;
+typedef int24_t int_least24_t;
+typedef uint24_t uint_least24_t;
+typedef int24_t int_fast24_t;
+typedef uint24_t uint_fast24_t;
+# define __int_least16_t int24_t
+# define __uint_least16_t uint24_t
+# define __int_least8_t int24_t
+# define __uint_least8_t uint24_t
+#endif /* __INT24_TYPE__ */
+
+#ifdef __INT16_TYPE__
+#ifndef __int8_t_defined /* glibc sys/types.h also defines int16_t*/
+typedef signed __INT16_TYPE__ int16_t;
+#endif /* __int8_t_defined */
+typedef unsigned __INT16_TYPE__ uint16_t;
+# define __int_least16_t int16_t
+# define __uint_least16_t uint16_t
+# define __int_least8_t int16_t
+# define __uint_least8_t uint16_t
+#endif /* __INT16_TYPE__ */
+
+#ifdef __int_least16_t
+typedef __int_least16_t int_least16_t;
+typedef __uint_least16_t uint_least16_t;
+typedef __int_least16_t int_fast16_t;
+typedef __uint_least16_t uint_fast16_t;
+#endif /* __int_least16_t */
+
+
+#ifdef __INT8_TYPE__
+#ifndef __int8_t_defined  /* glibc sys/types.h also defines int8_t*/
+typedef signed __INT8_TYPE__ int8_t;
+#endif /* __int8_t_defined */
+typedef unsigned __INT8_TYPE__ uint8_t;
+# define __int_least8_t int8_t
+# define __uint_least8_t uint8_t
+#endif /* __INT8_TYPE__ */
+
+#ifdef __int_least8_t
+typedef __int_least8_t int_least8_t;
+typedef __uint_least8_t uint_least8_t;
+typedef __int_least8_t int_fast8_t;
+typedef __uint_least8_t uint_fast8_t;
+#endif /* __int_least8_t */
+
+/* prevent glibc sys/types.h from defining conflicting types */
+#ifndef __int8_t_defined  
+# define __int8_t_defined
+#endif /* __int8_t_defined */
+
+/* C99 7.18.1.4 Integer types capable of holding object pointers.
+ */
+#define __stdint_join3(a,b,c) a ## b ## c
+
+#define  __intn_t(n) __stdint_join3( int, n, _t)
+#define __uintn_t(n) __stdint_join3(uint, n, _t)
+
+#ifndef _INTPTR_T
+#ifndef __intptr_t_defined
+typedef  __intn_t(__INTPTR_WIDTH__)  intptr_t;
+#define __intptr_t_defined
+#define _INTPTR_T
+#endif
+#endif
+
+#ifndef _UINTPTR_T
+typedef __uintn_t(__INTPTR_WIDTH__) uintptr_t;
+#define _UINTPTR_T
+#endif
+
+/* C99 7.18.1.5 Greatest-width integer types.
+ */
+typedef __INTMAX_TYPE__  intmax_t;
+typedef __UINTMAX_TYPE__ uintmax_t;
+
+/* C99 7.18.4 Macros for minimum-width integer constants.
+ *
+ * The standard requires that integer constant macros be defined for all the
+ * minimum-width types defined above. As 8-, 16-, 32-, and 64-bit minimum-width
+ * types are required, the corresponding integer constant macros are defined 
+ * here. This implementation also defines minimum-width types for every other
+ * integer width that the target implements, so corresponding macros are 
+ * defined below, too.
+ *
+ * These macros are defined using the same successive-shrinking approach as
+ * the type definitions above. It is likewise important that macros are defined
+ * in order of decending width.
+ *
+ * Note that C++ should not check __STDC_CONSTANT_MACROS here, contrary to the
+ * claims of the C standard (see C++ 18.3.1p2, [cstdint.syn]).
+ */
+
+#define __int_c_join(a, b) a ## b
+#define __int_c(v, suffix) __int_c_join(v, suffix)
+#define __uint_c(v, suffix) __int_c_join(v##U, suffix)
+
+
+#ifdef __INT64_TYPE__
+# ifdef __INT64_C_SUFFIX__
+#  define __int64_c_suffix __INT64_C_SUFFIX__
+#  define __int32_c_suffix __INT64_C_SUFFIX__
+#  define __int16_c_suffix __INT64_C_SUFFIX__
+#  define  __int8_c_suffix __INT64_C_SUFFIX__
+# else
+#  undef __int64_c_suffix
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT64_C_SUFFIX__ */
+#endif /* __INT64_TYPE__ */
+
+#ifdef __int_least64_t
+# ifdef __int64_c_suffix
+#  define INT64_C(v) __int_c(v, __int64_c_suffix)
+#  define UINT64_C(v) __uint_c(v, __int64_c_suffix)
+# else
+#  define INT64_C(v) v
+#  define UINT64_C(v) v ## U
+# endif /* __int64_c_suffix */
+#endif /* __int_least64_t */
+
+
+#ifdef __INT56_TYPE__
+# ifdef __INT56_C_SUFFIX__
+#  define INT56_C(v) __int_c(v, __INT56_C_SUFFIX__)
+#  define UINT56_C(v) __uint_c(v, __INT56_C_SUFFIX__)
+#  define __int32_c_suffix __INT56_C_SUFFIX__
+#  define __int16_c_suffix __INT56_C_SUFFIX__
+#  define __int8_c_suffix  __INT56_C_SUFFIX__
+# else
+#  define INT56_C(v) v
+#  define UINT56_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT56_C_SUFFIX__ */
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+# ifdef __INT48_C_SUFFIX__
+#  define INT48_C(v) __int_c(v, __INT48_C_SUFFIX__)
+#  define UINT48_C(v) __uint_c(v, __INT48_C_SUFFIX__)
+#  define __int32_c_suffix __INT48_C_SUFFIX__
+#  define __int16_c_suffix __INT48_C_SUFFIX__
+#  define __int8_c_suffix  __INT48_C_SUFFIX__
+# else
+#  define INT48_C(v) v
+#  define UINT48_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT48_C_SUFFIX__ */
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+# ifdef __INT40_C_SUFFIX__
+#  define INT40_C(v) __int_c(v, __INT40_C_SUFFIX__)
+#  define UINT40_C(v) __uint_c(v, __INT40_C_SUFFIX__)
+#  define __int32_c_suffix __INT40_C_SUFFIX__
+#  define __int16_c_suffix __INT40_C_SUFFIX__
+#  define __int8_c_suffix  __INT40_C_SUFFIX__
+# else
+#  define INT40_C(v) v
+#  define UINT40_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT40_C_SUFFIX__ */
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+# ifdef __INT32_C_SUFFIX__
+#  define __int32_c_suffix __INT32_C_SUFFIX__
+#  define __int16_c_suffix __INT32_C_SUFFIX__
+#  define __int8_c_suffix  __INT32_C_SUFFIX__
+#else
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT32_C_SUFFIX__ */
+#endif /* __INT32_TYPE__ */
+
+#ifdef __int_least32_t
+# ifdef __int32_c_suffix
+#  define INT32_C(v) __int_c(v, __int32_c_suffix)
+#  define UINT32_C(v) __uint_c(v, __int32_c_suffix)
+# else
+#  define INT32_C(v) v
+#  define UINT32_C(v) v ## U
+# endif /* __int32_c_suffix */
+#endif /* __int_least32_t */
+
+
+#ifdef __INT24_TYPE__
+# ifdef __INT24_C_SUFFIX__
+#  define INT24_C(v) __int_c(v, __INT24_C_SUFFIX__)
+#  define UINT24_C(v) __uint_c(v, __INT24_C_SUFFIX__)
+#  define __int16_c_suffix __INT24_C_SUFFIX__
+#  define __int8_c_suffix  __INT24_C_SUFFIX__
+# else
+#  define INT24_C(v) v
+#  define UINT24_C(v) v ## U
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT24_C_SUFFIX__ */
+#endif /* __INT24_TYPE__ */
+
+
+#ifdef __INT16_TYPE__
+# ifdef __INT16_C_SUFFIX__
+#  define __int16_c_suffix __INT16_C_SUFFIX__
+#  define __int8_c_suffix  __INT16_C_SUFFIX__
+#else
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT16_C_SUFFIX__ */
+#endif /* __INT16_TYPE__ */
+
+#ifdef __int_least16_t
+# ifdef __int16_c_suffix
+#  define INT16_C(v) __int_c(v, __int16_c_suffix)
+#  define UINT16_C(v) __uint_c(v, __int16_c_suffix)
+# else
+#  define INT16_C(v) v
+#  define UINT16_C(v) v ## U
+# endif /* __int16_c_suffix */
+#endif /* __int_least16_t */
+
+
+#ifdef __INT8_TYPE__
+# ifdef __INT8_C_SUFFIX__
+#  define __int8_c_suffix __INT8_C_SUFFIX__
+#else
+#  undef  __int8_c_suffix
+# endif /* __INT8_C_SUFFIX__ */
+#endif /* __INT8_TYPE__ */
+
+#ifdef __int_least8_t
+# ifdef __int8_c_suffix
+#  define INT8_C(v) __int_c(v, __int8_c_suffix)
+#  define UINT8_C(v) __uint_c(v, __int8_c_suffix)
+# else
+#  define INT8_C(v) v
+#  define UINT8_C(v) v ## U
+# endif /* __int8_c_suffix */
+#endif /* __int_least8_t */
+
+
+/* C99 7.18.2.1 Limits of exact-width integer types. 
+ * C99 7.18.2.2 Limits of minimum-width integer types.
+ * C99 7.18.2.3 Limits of fastest minimum-width integer types.
+ *
+ * The presence of limit macros are completely optional in C99.  This
+ * implementation defines limits for all of the types (exact- and
+ * minimum-width) that it defines above, using the limits of the minimum-width
+ * type for any types that do not have exact-width representations.
+ *
+ * As in the type definitions, this section takes an approach of
+ * successive-shrinking to determine which limits to use for the standard (8,
+ * 16, 32, 64) bit widths when they don't have exact representations. It is
+ * therefore important that the defintions be kept in order of decending
+ * widths.
+ *
+ * Note that C++ should not check __STDC_LIMIT_MACROS here, contrary to the
+ * claims of the C standard (see C++ 18.3.1p2, [cstdint.syn]).
+ */
+
+#ifdef __INT64_TYPE__
+# define INT64_MAX           INT64_C( 9223372036854775807)
+# define INT64_MIN         (-INT64_C( 9223372036854775807)-1)
+# define UINT64_MAX         UINT64_C(18446744073709551615)
+# define __INT_LEAST64_MIN   INT64_MIN
+# define __INT_LEAST64_MAX   INT64_MAX
+# define __UINT_LEAST64_MAX UINT64_MAX
+# define __INT_LEAST32_MIN   INT64_MIN
+# define __INT_LEAST32_MAX   INT64_MAX
+# define __UINT_LEAST32_MAX UINT64_MAX
+# define __INT_LEAST16_MIN   INT64_MIN
+# define __INT_LEAST16_MAX   INT64_MAX
+# define __UINT_LEAST16_MAX UINT64_MAX
+# define __INT_LEAST8_MIN    INT64_MIN
+# define __INT_LEAST8_MAX    INT64_MAX
+# define __UINT_LEAST8_MAX  UINT64_MAX
+#endif /* __INT64_TYPE__ */
+
+#ifdef __INT_LEAST64_MIN
+# define INT_LEAST64_MIN   __INT_LEAST64_MIN
+# define INT_LEAST64_MAX   __INT_LEAST64_MAX
+# define UINT_LEAST64_MAX __UINT_LEAST64_MAX
+# define INT_FAST64_MIN    __INT_LEAST64_MIN
+# define INT_FAST64_MAX    __INT_LEAST64_MAX
+# define UINT_FAST64_MAX  __UINT_LEAST64_MAX
+#endif /* __INT_LEAST64_MIN */
+
+
+#ifdef __INT56_TYPE__
+# define INT56_MAX           INT56_C(36028797018963967)
+# define INT56_MIN         (-INT56_C(36028797018963967)-1)
+# define UINT56_MAX         UINT56_C(72057594037927935)
+# define INT_LEAST56_MIN     INT56_MIN
+# define INT_LEAST56_MAX     INT56_MAX
+# define UINT_LEAST56_MAX   UINT56_MAX
+# define INT_FAST56_MIN      INT56_MIN
+# define INT_FAST56_MAX      INT56_MAX
+# define UINT_FAST56_MAX    UINT56_MAX
+# define __INT_LEAST32_MIN   INT56_MIN
+# define __INT_LEAST32_MAX   INT56_MAX
+# define __UINT_LEAST32_MAX UINT56_MAX
+# define __INT_LEAST16_MIN   INT56_MIN
+# define __INT_LEAST16_MAX   INT56_MAX
+# define __UINT_LEAST16_MAX UINT56_MAX
+# define __INT_LEAST8_MIN    INT56_MIN
+# define __INT_LEAST8_MAX    INT56_MAX
+# define __UINT_LEAST8_MAX  UINT56_MAX
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+# define INT48_MAX           INT48_C(140737488355327)
+# define INT48_MIN         (-INT48_C(140737488355327)-1)
+# define UINT48_MAX         UINT48_C(281474976710655)
+# define INT_LEAST48_MIN     INT48_MIN
+# define INT_LEAST48_MAX     INT48_MAX
+# define UINT_LEAST48_MAX   UINT48_MAX
+# define INT_FAST48_MIN      INT48_MIN
+# define INT_FAST48_MAX      INT48_MAX
+# define UINT_FAST48_MAX    UINT48_MAX
+# define __INT_LEAST32_MIN   INT48_MIN
+# define __INT_LEAST32_MAX   INT48_MAX
+# define __UINT_LEAST32_MAX UINT48_MAX
+# define __INT_LEAST16_MIN   INT48_MIN
+# define __INT_LEAST16_MAX   INT48_MAX
+# define __UINT_LEAST16_MAX UINT48_MAX
+# define __INT_LEAST8_MIN    INT48_MIN
+# define __INT_LEAST8_MAX    INT48_MAX
+# define __UINT_LEAST8_MAX  UINT48_MAX
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+# define INT40_MAX           INT40_C(549755813887)
+# define INT40_MIN         (-INT40_C(549755813887)-1)
+# define UINT40_MAX         UINT40_C(1099511627775)
+# define INT_LEAST40_MIN     INT40_MIN
+# define INT_LEAST40_MAX     INT40_MAX
+# define UINT_LEAST40_MAX   UINT40_MAX
+# define INT_FAST40_MIN      INT40_MIN
+# define INT_FAST40_MAX      INT40_MAX
+# define UINT_FAST40_MAX    UINT40_MAX
+# define __INT_LEAST32_MIN   INT40_MIN
+# define __INT_LEAST32_MAX   INT40_MAX
+# define __UINT_LEAST32_MAX UINT40_MAX
+# define __INT_LEAST16_MIN   INT40_MIN
+# define __INT_LEAST16_MAX   INT40_MAX
+# define __UINT_LEAST16_MAX UINT40_MAX
+# define __INT_LEAST8_MIN    INT40_MIN
+# define __INT_LEAST8_MAX    INT40_MAX
+# define __UINT_LEAST8_MAX  UINT40_MAX
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+# define INT32_MAX           INT32_C(2147483647)
+# define INT32_MIN         (-INT32_C(2147483647)-1)
+# define UINT32_MAX         UINT32_C(4294967295)
+# define __INT_LEAST32_MIN   INT32_MIN
+# define __INT_LEAST32_MAX   INT32_MAX
+# define __UINT_LEAST32_MAX UINT32_MAX
+# define __INT_LEAST16_MIN   INT32_MIN
+# define __INT_LEAST16_MAX   INT32_MAX
+# define __UINT_LEAST16_MAX UINT32_MAX
+# define __INT_LEAST8_MIN    INT32_MIN
+# define __INT_LEAST8_MAX    INT32_MAX
+# define __UINT_LEAST8_MAX  UINT32_MAX
+#endif /* __INT32_TYPE__ */
+
+#ifdef __INT_LEAST32_MIN
+# define INT_LEAST32_MIN   __INT_LEAST32_MIN
+# define INT_LEAST32_MAX   __INT_LEAST32_MAX
+# define UINT_LEAST32_MAX __UINT_LEAST32_MAX
+# define INT_FAST32_MIN    __INT_LEAST32_MIN
+# define INT_FAST32_MAX    __INT_LEAST32_MAX
+# define UINT_FAST32_MAX  __UINT_LEAST32_MAX
+#endif /* __INT_LEAST32_MIN */
+
+
+#ifdef __INT24_TYPE__
+# define INT24_MAX           INT24_C(8388607)
+# define INT24_MIN         (-INT24_C(8388607)-1)
+# define UINT24_MAX         UINT24_C(16777215)
+# define INT_LEAST24_MIN     INT24_MIN
+# define INT_LEAST24_MAX     INT24_MAX
+# define UINT_LEAST24_MAX   UINT24_MAX
+# define INT_FAST24_MIN      INT24_MIN
+# define INT_FAST24_MAX      INT24_MAX
+# define UINT_FAST24_MAX    UINT24_MAX
+# define __INT_LEAST16_MIN   INT24_MIN
+# define __INT_LEAST16_MAX   INT24_MAX
+# define __UINT_LEAST16_MAX UINT24_MAX
+# define __INT_LEAST8_MIN    INT24_MIN
+# define __INT_LEAST8_MAX    INT24_MAX
+# define __UINT_LEAST8_MAX  UINT24_MAX
+#endif /* __INT24_TYPE__ */
+
+
+#ifdef __INT16_TYPE__
+#define INT16_MAX            INT16_C(32767)
+#define INT16_MIN          (-INT16_C(32767)-1)
+#define UINT16_MAX          UINT16_C(65535)
+# define __INT_LEAST16_MIN   INT16_MIN
+# define __INT_LEAST16_MAX   INT16_MAX
+# define __UINT_LEAST16_MAX UINT16_MAX
+# define __INT_LEAST8_MIN    INT16_MIN
+# define __INT_LEAST8_MAX    INT16_MAX
+# define __UINT_LEAST8_MAX  UINT16_MAX
+#endif /* __INT16_TYPE__ */
+
+#ifdef __INT_LEAST16_MIN
+# define INT_LEAST16_MIN   __INT_LEAST16_MIN
+# define INT_LEAST16_MAX   __INT_LEAST16_MAX
+# define UINT_LEAST16_MAX __UINT_LEAST16_MAX
+# define INT_FAST16_MIN    __INT_LEAST16_MIN
+# define INT_FAST16_MAX    __INT_LEAST16_MAX
+# define UINT_FAST16_MAX  __UINT_LEAST16_MAX
+#endif /* __INT_LEAST16_MIN */
+
+
+#ifdef __INT8_TYPE__
+# define INT8_MAX            INT8_C(127)
+# define INT8_MIN          (-INT8_C(127)-1)
+# define UINT8_MAX          UINT8_C(255)
+# define __INT_LEAST8_MIN    INT8_MIN
+# define __INT_LEAST8_MAX    INT8_MAX
+# define __UINT_LEAST8_MAX  UINT8_MAX
+#endif /* __INT8_TYPE__ */
+
+#ifdef __INT_LEAST8_MIN
+# define INT_LEAST8_MIN   __INT_LEAST8_MIN
+# define INT_LEAST8_MAX   __INT_LEAST8_MAX
+# define UINT_LEAST8_MAX __UINT_LEAST8_MAX
+# define INT_FAST8_MIN    __INT_LEAST8_MIN
+# define INT_FAST8_MAX    __INT_LEAST8_MAX
+# define UINT_FAST8_MAX  __UINT_LEAST8_MAX
+#endif /* __INT_LEAST8_MIN */
+
+/* Some utility macros */
+#define  __INTN_MIN(n)  __stdint_join3( INT, n, _MIN)
+#define  __INTN_MAX(n)  __stdint_join3( INT, n, _MAX)
+#define __UINTN_MAX(n)  __stdint_join3(UINT, n, _MAX)
+#define  __INTN_C(n, v) __stdint_join3( INT, n, _C(v))
+#define __UINTN_C(n, v) __stdint_join3(UINT, n, _C(v))
+
+/* C99 7.18.2.4 Limits of integer types capable of holding object pointers. */
+/* C99 7.18.3 Limits of other integer types. */
+
+#define  INTPTR_MIN  __INTN_MIN(__INTPTR_WIDTH__)
+#define  INTPTR_MAX  __INTN_MAX(__INTPTR_WIDTH__)
+#define UINTPTR_MAX __UINTN_MAX(__INTPTR_WIDTH__)
+#define PTRDIFF_MIN  __INTN_MIN(__PTRDIFF_WIDTH__)
+#define PTRDIFF_MAX  __INTN_MAX(__PTRDIFF_WIDTH__)
+#define    SIZE_MAX __UINTN_MAX(__SIZE_WIDTH__)
+
+/* C99 7.18.2.5 Limits of greatest-width integer types. */
+#define INTMAX_MIN   __INTN_MIN(__INTMAX_WIDTH__)
+#define INTMAX_MAX   __INTN_MAX(__INTMAX_WIDTH__)
+#define UINTMAX_MAX __UINTN_MAX(__INTMAX_WIDTH__)
+
+/* C99 7.18.3 Limits of other integer types. */
+#define SIG_ATOMIC_MIN __INTN_MIN(__SIG_ATOMIC_WIDTH__)
+#define SIG_ATOMIC_MAX __INTN_MAX(__SIG_ATOMIC_WIDTH__)
+#ifdef __WINT_UNSIGNED__
+# define WINT_MIN       __UINTN_C(__WINT_WIDTH__, 0)
+# define WINT_MAX       __UINTN_MAX(__WINT_WIDTH__)
+#else
+# define WINT_MIN       __INTN_MIN(__WINT_WIDTH__)
+# define WINT_MAX       __INTN_MAX(__WINT_WIDTH__)
+#endif
+
+#ifndef WCHAR_MAX
+# define WCHAR_MAX __WCHAR_MAX__
+#endif
+#ifndef WCHAR_MIN
+# if __WCHAR_MAX__ == __INTN_MAX(__WCHAR_WIDTH__)
+#  define WCHAR_MIN __INTN_MIN(__WCHAR_WIDTH__)
+# else
+#  define WCHAR_MIN __UINTN_C(__WCHAR_WIDTH__, 0)
+# endif
+#endif
+
+/* 7.18.4.2 Macros for greatest-width integer constants. */
+#define INTMAX_C(v)   __INTN_C(__INTMAX_WIDTH__, v)
+#define UINTMAX_C(v) __UINTN_C(__INTMAX_WIDTH__, v)
+
+#endif /* __STDC_HOSTED__ */
+#endif /* __CLANG_STDINT_H */
diff --git a/clang/lib/Headers/tgmath.h b/clang/lib/Headers/tgmath.h
new file mode 100644
index 0000000..4fa1cf7
--- /dev/null
+++ b/clang/lib/Headers/tgmath.h
@@ -0,0 +1,1374 @@
+/*===---- tgmath.h - Standard header for type generic math ----------------===*\
+ *
+ * Copyright (c) 2009 Howard Hinnant
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __TGMATH_H
+#define __TGMATH_H
+
+/* C99 7.22 Type-generic math <tgmath.h>. */
+#include <math.h>
+
+/* C++ handles type genericity with overloading in math.h. */
+#ifndef __cplusplus
+#include <complex.h>
+
+#define _TG_ATTRSp __attribute__((__overloadable__))
+#define _TG_ATTRS __attribute__((__overloadable__, __always_inline__))
+
+// promotion
+
+typedef void _Argument_type_is_not_arithmetic;
+static _Argument_type_is_not_arithmetic __tg_promote(...)
+  __attribute__((__unavailable__,__overloadable__));
+static double               _TG_ATTRSp __tg_promote(int);
+static double               _TG_ATTRSp __tg_promote(unsigned int);
+static double               _TG_ATTRSp __tg_promote(long);
+static double               _TG_ATTRSp __tg_promote(unsigned long);
+static double               _TG_ATTRSp __tg_promote(long long);
+static double               _TG_ATTRSp __tg_promote(unsigned long long);
+static float                _TG_ATTRSp __tg_promote(float);
+static double               _TG_ATTRSp __tg_promote(double);
+static long double          _TG_ATTRSp __tg_promote(long double);
+static float _Complex       _TG_ATTRSp __tg_promote(float _Complex);
+static double _Complex      _TG_ATTRSp __tg_promote(double _Complex);
+static long double _Complex _TG_ATTRSp __tg_promote(long double _Complex);
+
+#define __tg_promote1(__x)           (__typeof__(__tg_promote(__x)))
+#define __tg_promote2(__x, __y)      (__typeof__(__tg_promote(__x) + \
+                                                 __tg_promote(__y)))
+#define __tg_promote3(__x, __y, __z) (__typeof__(__tg_promote(__x) + \
+                                                 __tg_promote(__y) + \
+                                                 __tg_promote(__z)))
+
+// acos
+
+static float
+    _TG_ATTRS
+    __tg_acos(float __x) {return acosf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_acos(double __x) {return acos(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_acos(long double __x) {return acosl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_acos(float _Complex __x) {return cacosf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_acos(double _Complex __x) {return cacos(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_acos(long double _Complex __x) {return cacosl(__x);}
+
+#undef acos
+#define acos(__x) __tg_acos(__tg_promote1((__x))(__x))
+
+// asin
+
+static float
+    _TG_ATTRS
+    __tg_asin(float __x) {return asinf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_asin(double __x) {return asin(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_asin(long double __x) {return asinl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_asin(float _Complex __x) {return casinf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_asin(double _Complex __x) {return casin(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_asin(long double _Complex __x) {return casinl(__x);}
+
+#undef asin
+#define asin(__x) __tg_asin(__tg_promote1((__x))(__x))
+
+// atan
+
+static float
+    _TG_ATTRS
+    __tg_atan(float __x) {return atanf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_atan(double __x) {return atan(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_atan(long double __x) {return atanl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_atan(float _Complex __x) {return catanf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_atan(double _Complex __x) {return catan(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_atan(long double _Complex __x) {return catanl(__x);}
+
+#undef atan
+#define atan(__x) __tg_atan(__tg_promote1((__x))(__x))
+
+// acosh
+
+static float
+    _TG_ATTRS
+    __tg_acosh(float __x) {return acoshf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_acosh(double __x) {return acosh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_acosh(long double __x) {return acoshl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_acosh(float _Complex __x) {return cacoshf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_acosh(double _Complex __x) {return cacosh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_acosh(long double _Complex __x) {return cacoshl(__x);}
+
+#undef acosh
+#define acosh(__x) __tg_acosh(__tg_promote1((__x))(__x))
+
+// asinh
+
+static float
+    _TG_ATTRS
+    __tg_asinh(float __x) {return asinhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_asinh(double __x) {return asinh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_asinh(long double __x) {return asinhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_asinh(float _Complex __x) {return casinhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_asinh(double _Complex __x) {return casinh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_asinh(long double _Complex __x) {return casinhl(__x);}
+
+#undef asinh
+#define asinh(__x) __tg_asinh(__tg_promote1((__x))(__x))
+
+// atanh
+
+static float
+    _TG_ATTRS
+    __tg_atanh(float __x) {return atanhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_atanh(double __x) {return atanh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_atanh(long double __x) {return atanhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_atanh(float _Complex __x) {return catanhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_atanh(double _Complex __x) {return catanh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_atanh(long double _Complex __x) {return catanhl(__x);}
+
+#undef atanh
+#define atanh(__x) __tg_atanh(__tg_promote1((__x))(__x))
+
+// cos
+
+static float
+    _TG_ATTRS
+    __tg_cos(float __x) {return cosf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cos(double __x) {return cos(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cos(long double __x) {return cosl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cos(float _Complex __x) {return ccosf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cos(double _Complex __x) {return ccos(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cos(long double _Complex __x) {return ccosl(__x);}
+
+#undef cos
+#define cos(__x) __tg_cos(__tg_promote1((__x))(__x))
+
+// sin
+
+static float
+    _TG_ATTRS
+    __tg_sin(float __x) {return sinf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sin(double __x) {return sin(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sin(long double __x) {return sinl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sin(float _Complex __x) {return csinf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sin(double _Complex __x) {return csin(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sin(long double _Complex __x) {return csinl(__x);}
+
+#undef sin
+#define sin(__x) __tg_sin(__tg_promote1((__x))(__x))
+
+// tan
+
+static float
+    _TG_ATTRS
+    __tg_tan(float __x) {return tanf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tan(double __x) {return tan(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tan(long double __x) {return tanl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_tan(float _Complex __x) {return ctanf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_tan(double _Complex __x) {return ctan(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_tan(long double _Complex __x) {return ctanl(__x);}
+
+#undef tan
+#define tan(__x) __tg_tan(__tg_promote1((__x))(__x))
+
+// cosh
+
+static float
+    _TG_ATTRS
+    __tg_cosh(float __x) {return coshf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cosh(double __x) {return cosh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cosh(long double __x) {return coshl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cosh(float _Complex __x) {return ccoshf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cosh(double _Complex __x) {return ccosh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cosh(long double _Complex __x) {return ccoshl(__x);}
+
+#undef cosh
+#define cosh(__x) __tg_cosh(__tg_promote1((__x))(__x))
+
+// sinh
+
+static float
+    _TG_ATTRS
+    __tg_sinh(float __x) {return sinhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sinh(double __x) {return sinh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sinh(long double __x) {return sinhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sinh(float _Complex __x) {return csinhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sinh(double _Complex __x) {return csinh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sinh(long double _Complex __x) {return csinhl(__x);}
+
+#undef sinh
+#define sinh(__x) __tg_sinh(__tg_promote1((__x))(__x))
+
+// tanh
+
+static float
+    _TG_ATTRS
+    __tg_tanh(float __x) {return tanhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tanh(double __x) {return tanh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tanh(long double __x) {return tanhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_tanh(float _Complex __x) {return ctanhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_tanh(double _Complex __x) {return ctanh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_tanh(long double _Complex __x) {return ctanhl(__x);}
+
+#undef tanh
+#define tanh(__x) __tg_tanh(__tg_promote1((__x))(__x))
+
+// exp
+
+static float
+    _TG_ATTRS
+    __tg_exp(float __x) {return expf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_exp(double __x) {return exp(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_exp(long double __x) {return expl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_exp(float _Complex __x) {return cexpf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_exp(double _Complex __x) {return cexp(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_exp(long double _Complex __x) {return cexpl(__x);}
+
+#undef exp
+#define exp(__x) __tg_exp(__tg_promote1((__x))(__x))
+
+// log
+
+static float
+    _TG_ATTRS
+    __tg_log(float __x) {return logf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log(double __x) {return log(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log(long double __x) {return logl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_log(float _Complex __x) {return clogf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_log(double _Complex __x) {return clog(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_log(long double _Complex __x) {return clogl(__x);}
+
+#undef log
+#define log(__x) __tg_log(__tg_promote1((__x))(__x))
+
+// pow
+
+static float
+    _TG_ATTRS
+    __tg_pow(float __x, float __y) {return powf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_pow(double __x, double __y) {return pow(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_pow(long double __x, long double __y) {return powl(__x, __y);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_pow(float _Complex __x, float _Complex __y) {return cpowf(__x, __y);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_pow(double _Complex __x, double _Complex __y) {return cpow(__x, __y);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_pow(long double _Complex __x, long double _Complex __y) 
+    {return cpowl(__x, __y);}
+
+#undef pow
+#define pow(__x, __y) __tg_pow(__tg_promote2((__x), (__y))(__x), \
+                               __tg_promote2((__x), (__y))(__y))
+
+// sqrt
+
+static float
+    _TG_ATTRS
+    __tg_sqrt(float __x) {return sqrtf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sqrt(double __x) {return sqrt(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sqrt(long double __x) {return sqrtl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sqrt(float _Complex __x) {return csqrtf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sqrt(double _Complex __x) {return csqrt(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sqrt(long double _Complex __x) {return csqrtl(__x);}
+
+#undef sqrt
+#define sqrt(__x) __tg_sqrt(__tg_promote1((__x))(__x))
+
+// fabs
+
+static float
+    _TG_ATTRS
+    __tg_fabs(float __x) {return fabsf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_fabs(double __x) {return fabs(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_fabs(long double __x) {return fabsl(__x);}
+
+static float
+    _TG_ATTRS
+    __tg_fabs(float _Complex __x) {return cabsf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_fabs(double _Complex __x) {return cabs(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_fabs(long double _Complex __x) {return cabsl(__x);}
+
+#undef fabs
+#define fabs(__x) __tg_fabs(__tg_promote1((__x))(__x))
+
+// atan2
+
+static float
+    _TG_ATTRS
+    __tg_atan2(float __x, float __y) {return atan2f(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_atan2(double __x, double __y) {return atan2(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_atan2(long double __x, long double __y) {return atan2l(__x, __y);}
+
+#undef atan2
+#define atan2(__x, __y) __tg_atan2(__tg_promote2((__x), (__y))(__x), \
+                                   __tg_promote2((__x), (__y))(__y))
+
+// cbrt
+
+static float
+    _TG_ATTRS
+    __tg_cbrt(float __x) {return cbrtf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cbrt(double __x) {return cbrt(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cbrt(long double __x) {return cbrtl(__x);}
+
+#undef cbrt
+#define cbrt(__x) __tg_cbrt(__tg_promote1((__x))(__x))
+
+// ceil
+
+static float
+    _TG_ATTRS
+    __tg_ceil(float __x) {return ceilf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_ceil(double __x) {return ceil(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_ceil(long double __x) {return ceill(__x);}
+
+#undef ceil
+#define ceil(__x) __tg_ceil(__tg_promote1((__x))(__x))
+
+// copysign
+
+static float
+    _TG_ATTRS
+    __tg_copysign(float __x, float __y) {return copysignf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_copysign(double __x, double __y) {return copysign(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_copysign(long double __x, long double __y) {return copysignl(__x, __y);}
+
+#undef copysign
+#define copysign(__x, __y) __tg_copysign(__tg_promote2((__x), (__y))(__x), \
+                                         __tg_promote2((__x), (__y))(__y))
+
+// erf
+
+static float
+    _TG_ATTRS
+    __tg_erf(float __x) {return erff(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_erf(double __x) {return erf(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_erf(long double __x) {return erfl(__x);}
+
+#undef erf
+#define erf(__x) __tg_erf(__tg_promote1((__x))(__x))
+
+// erfc
+
+static float
+    _TG_ATTRS
+    __tg_erfc(float __x) {return erfcf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_erfc(double __x) {return erfc(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_erfc(long double __x) {return erfcl(__x);}
+
+#undef erfc
+#define erfc(__x) __tg_erfc(__tg_promote1((__x))(__x))
+
+// exp2
+
+static float
+    _TG_ATTRS
+    __tg_exp2(float __x) {return exp2f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_exp2(double __x) {return exp2(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_exp2(long double __x) {return exp2l(__x);}
+
+#undef exp2
+#define exp2(__x) __tg_exp2(__tg_promote1((__x))(__x))
+
+// expm1
+
+static float
+    _TG_ATTRS
+    __tg_expm1(float __x) {return expm1f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_expm1(double __x) {return expm1(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_expm1(long double __x) {return expm1l(__x);}
+
+#undef expm1
+#define expm1(__x) __tg_expm1(__tg_promote1((__x))(__x))
+
+// fdim
+
+static float
+    _TG_ATTRS
+    __tg_fdim(float __x, float __y) {return fdimf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fdim(double __x, double __y) {return fdim(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fdim(long double __x, long double __y) {return fdiml(__x, __y);}
+
+#undef fdim
+#define fdim(__x, __y) __tg_fdim(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// floor
+
+static float
+    _TG_ATTRS
+    __tg_floor(float __x) {return floorf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_floor(double __x) {return floor(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_floor(long double __x) {return floorl(__x);}
+
+#undef floor
+#define floor(__x) __tg_floor(__tg_promote1((__x))(__x))
+
+// fma
+
+static float
+    _TG_ATTRS
+    __tg_fma(float __x, float __y, float __z)
+    {return fmaf(__x, __y, __z);}
+
+static double
+    _TG_ATTRS
+    __tg_fma(double __x, double __y, double __z)
+    {return fma(__x, __y, __z);}
+
+static long double
+    _TG_ATTRS
+    __tg_fma(long double __x,long double __y, long double __z)
+    {return fmal(__x, __y, __z);}
+
+#undef fma
+#define fma(__x, __y, __z)                                \
+        __tg_fma(__tg_promote3((__x), (__y), (__z))(__x), \
+                 __tg_promote3((__x), (__y), (__z))(__y), \
+                 __tg_promote3((__x), (__y), (__z))(__z))
+
+// fmax
+
+static float
+    _TG_ATTRS
+    __tg_fmax(float __x, float __y) {return fmaxf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmax(double __x, double __y) {return fmax(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmax(long double __x, long double __y) {return fmaxl(__x, __y);}
+
+#undef fmax
+#define fmax(__x, __y) __tg_fmax(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// fmin
+
+static float
+    _TG_ATTRS
+    __tg_fmin(float __x, float __y) {return fminf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmin(double __x, double __y) {return fmin(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmin(long double __x, long double __y) {return fminl(__x, __y);}
+
+#undef fmin
+#define fmin(__x, __y) __tg_fmin(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// fmod
+
+static float
+    _TG_ATTRS
+    __tg_fmod(float __x, float __y) {return fmodf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmod(double __x, double __y) {return fmod(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmod(long double __x, long double __y) {return fmodl(__x, __y);}
+
+#undef fmod
+#define fmod(__x, __y) __tg_fmod(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// frexp
+
+static float
+    _TG_ATTRS
+    __tg_frexp(float __x, int* __y) {return frexpf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_frexp(double __x, int* __y) {return frexp(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_frexp(long double __x, int* __y) {return frexpl(__x, __y);}
+
+#undef frexp
+#define frexp(__x, __y) __tg_frexp(__tg_promote1((__x))(__x), __y)
+
+// hypot
+
+static float
+    _TG_ATTRS
+    __tg_hypot(float __x, float __y) {return hypotf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_hypot(double __x, double __y) {return hypot(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_hypot(long double __x, long double __y) {return hypotl(__x, __y);}
+
+#undef hypot
+#define hypot(__x, __y) __tg_hypot(__tg_promote2((__x), (__y))(__x), \
+                                   __tg_promote2((__x), (__y))(__y))
+
+// ilogb
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(float __x) {return ilogbf(__x);}
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(double __x) {return ilogb(__x);}
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(long double __x) {return ilogbl(__x);}
+
+#undef ilogb
+#define ilogb(__x) __tg_ilogb(__tg_promote1((__x))(__x))
+
+// ldexp
+
+static float
+    _TG_ATTRS
+    __tg_ldexp(float __x, int __y) {return ldexpf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_ldexp(double __x, int __y) {return ldexp(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_ldexp(long double __x, int __y) {return ldexpl(__x, __y);}
+
+#undef ldexp
+#define ldexp(__x, __y) __tg_ldexp(__tg_promote1((__x))(__x), __y)
+
+// lgamma
+
+static float
+    _TG_ATTRS
+    __tg_lgamma(float __x) {return lgammaf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_lgamma(double __x) {return lgamma(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_lgamma(long double __x) {return lgammal(__x);}
+
+#undef lgamma
+#define lgamma(__x) __tg_lgamma(__tg_promote1((__x))(__x))
+
+// llrint
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(float __x) {return llrintf(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(double __x) {return llrint(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(long double __x) {return llrintl(__x);}
+
+#undef llrint
+#define llrint(__x) __tg_llrint(__tg_promote1((__x))(__x))
+
+// llround
+
+static long long
+    _TG_ATTRS
+    __tg_llround(float __x) {return llroundf(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llround(double __x) {return llround(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llround(long double __x) {return llroundl(__x);}
+
+#undef llround
+#define llround(__x) __tg_llround(__tg_promote1((__x))(__x))
+
+// log10
+
+static float
+    _TG_ATTRS
+    __tg_log10(float __x) {return log10f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log10(double __x) {return log10(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log10(long double __x) {return log10l(__x);}
+
+#undef log10
+#define log10(__x) __tg_log10(__tg_promote1((__x))(__x))
+
+// log1p
+
+static float
+    _TG_ATTRS
+    __tg_log1p(float __x) {return log1pf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log1p(double __x) {return log1p(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log1p(long double __x) {return log1pl(__x);}
+
+#undef log1p
+#define log1p(__x) __tg_log1p(__tg_promote1((__x))(__x))
+
+// log2
+
+static float
+    _TG_ATTRS
+    __tg_log2(float __x) {return log2f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log2(double __x) {return log2(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log2(long double __x) {return log2l(__x);}
+
+#undef log2
+#define log2(__x) __tg_log2(__tg_promote1((__x))(__x))
+
+// logb
+
+static float
+    _TG_ATTRS
+    __tg_logb(float __x) {return logbf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_logb(double __x) {return logb(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_logb(long double __x) {return logbl(__x);}
+
+#undef logb
+#define logb(__x) __tg_logb(__tg_promote1((__x))(__x))
+
+// lrint
+
+static long
+    _TG_ATTRS
+    __tg_lrint(float __x) {return lrintf(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lrint(double __x) {return lrint(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lrint(long double __x) {return lrintl(__x);}
+
+#undef lrint
+#define lrint(__x) __tg_lrint(__tg_promote1((__x))(__x))
+
+// lround
+
+static long
+    _TG_ATTRS
+    __tg_lround(float __x) {return lroundf(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lround(double __x) {return lround(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lround(long double __x) {return lroundl(__x);}
+
+#undef lround
+#define lround(__x) __tg_lround(__tg_promote1((__x))(__x))
+
+// nearbyint
+
+static float
+    _TG_ATTRS
+    __tg_nearbyint(float __x) {return nearbyintf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_nearbyint(double __x) {return nearbyint(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_nearbyint(long double __x) {return nearbyintl(__x);}
+
+#undef nearbyint
+#define nearbyint(__x) __tg_nearbyint(__tg_promote1((__x))(__x))
+
+// nextafter
+
+static float
+    _TG_ATTRS
+    __tg_nextafter(float __x, float __y) {return nextafterf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_nextafter(double __x, double __y) {return nextafter(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_nextafter(long double __x, long double __y) {return nextafterl(__x, __y);}
+
+#undef nextafter
+#define nextafter(__x, __y) __tg_nextafter(__tg_promote2((__x), (__y))(__x), \
+                                           __tg_promote2((__x), (__y))(__y))
+
+// nexttoward
+
+static float
+    _TG_ATTRS
+    __tg_nexttoward(float __x, long double __y) {return nexttowardf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_nexttoward(double __x, long double __y) {return nexttoward(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_nexttoward(long double __x, long double __y) {return nexttowardl(__x, __y);}
+
+#undef nexttoward
+#define nexttoward(__x, __y) __tg_nexttoward(__tg_promote1((__x))(__x), (__y))
+
+// remainder
+
+static float
+    _TG_ATTRS
+    __tg_remainder(float __x, float __y) {return remainderf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_remainder(double __x, double __y) {return remainder(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_remainder(long double __x, long double __y) {return remainderl(__x, __y);}
+
+#undef remainder
+#define remainder(__x, __y) __tg_remainder(__tg_promote2((__x), (__y))(__x), \
+                                           __tg_promote2((__x), (__y))(__y))
+
+// remquo
+
+static float
+    _TG_ATTRS
+    __tg_remquo(float __x, float __y, int* __z)
+    {return remquof(__x, __y, __z);}
+
+static double
+    _TG_ATTRS
+    __tg_remquo(double __x, double __y, int* __z)
+    {return remquo(__x, __y, __z);}
+
+static long double
+    _TG_ATTRS
+    __tg_remquo(long double __x,long double __y, int* __z)
+    {return remquol(__x, __y, __z);}
+
+#undef remquo
+#define remquo(__x, __y, __z)                         \
+        __tg_remquo(__tg_promote2((__x), (__y))(__x), \
+                    __tg_promote2((__x), (__y))(__y), \
+                    (__z))
+
+// rint
+
+static float
+    _TG_ATTRS
+    __tg_rint(float __x) {return rintf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_rint(double __x) {return rint(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_rint(long double __x) {return rintl(__x);}
+
+#undef rint
+#define rint(__x) __tg_rint(__tg_promote1((__x))(__x))
+
+// round
+
+static float
+    _TG_ATTRS
+    __tg_round(float __x) {return roundf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_round(double __x) {return round(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_round(long double __x) {return roundl(__x);}
+
+#undef round
+#define round(__x) __tg_round(__tg_promote1((__x))(__x))
+
+// scalbn
+
+static float
+    _TG_ATTRS
+    __tg_scalbn(float __x, int __y) {return scalbnf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_scalbn(double __x, int __y) {return scalbn(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_scalbn(long double __x, int __y) {return scalbnl(__x, __y);}
+
+#undef scalbn
+#define scalbn(__x, __y) __tg_scalbn(__tg_promote1((__x))(__x), __y)
+
+// scalbln
+
+static float
+    _TG_ATTRS
+    __tg_scalbln(float __x, long __y) {return scalblnf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_scalbln(double __x, long __y) {return scalbln(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_scalbln(long double __x, long __y) {return scalblnl(__x, __y);}
+
+#undef scalbln
+#define scalbln(__x, __y) __tg_scalbln(__tg_promote1((__x))(__x), __y)
+
+// tgamma
+
+static float
+    _TG_ATTRS
+    __tg_tgamma(float __x) {return tgammaf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tgamma(double __x) {return tgamma(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tgamma(long double __x) {return tgammal(__x);}
+
+#undef tgamma
+#define tgamma(__x) __tg_tgamma(__tg_promote1((__x))(__x))
+
+// trunc
+
+static float
+    _TG_ATTRS
+    __tg_trunc(float __x) {return truncf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_trunc(double __x) {return trunc(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_trunc(long double __x) {return truncl(__x);}
+
+#undef trunc
+#define trunc(__x) __tg_trunc(__tg_promote1((__x))(__x))
+
+// carg
+
+static float
+    _TG_ATTRS
+    __tg_carg(float __x) {return atan2f(0.F, __x);}
+
+static double
+    _TG_ATTRS
+    __tg_carg(double __x) {return atan2(0., __x);}
+
+static long double
+    _TG_ATTRS
+    __tg_carg(long double __x) {return atan2l(0.L, __x);}
+
+static float
+    _TG_ATTRS
+    __tg_carg(float _Complex __x) {return cargf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_carg(double _Complex __x) {return carg(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_carg(long double _Complex __x) {return cargl(__x);}
+
+#undef carg
+#define carg(__x) __tg_carg(__tg_promote1((__x))(__x))
+
+// cimag
+
+static float
+    _TG_ATTRS
+    __tg_cimag(float __x) {return 0;}
+
+static double
+    _TG_ATTRS
+    __tg_cimag(double __x) {return 0;}
+
+static long double
+    _TG_ATTRS
+    __tg_cimag(long double __x) {return 0;}
+
+static float
+    _TG_ATTRS
+    __tg_cimag(float _Complex __x) {return cimagf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cimag(double _Complex __x) {return cimag(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cimag(long double _Complex __x) {return cimagl(__x);}
+
+#undef cimag
+#define cimag(__x) __tg_cimag(__tg_promote1((__x))(__x))
+
+// conj
+
+static float _Complex
+    _TG_ATTRS
+    __tg_conj(float __x) {return __x;}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_conj(double __x) {return __x;}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_conj(long double __x) {return __x;}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_conj(float _Complex __x) {return conjf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_conj(double _Complex __x) {return conj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_conj(long double _Complex __x) {return conjl(__x);}
+
+#undef conj
+#define conj(__x) __tg_conj(__tg_promote1((__x))(__x))
+
+// cproj
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cproj(float __x) {return cprojf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cproj(double __x) {return cproj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cproj(long double __x) {return cprojl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cproj(float _Complex __x) {return cprojf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cproj(double _Complex __x) {return cproj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cproj(long double _Complex __x) {return cprojl(__x);}
+
+#undef cproj
+#define cproj(__x) __tg_cproj(__tg_promote1((__x))(__x))
+
+// creal
+
+static float _Complex
+    _TG_ATTRS
+    __tg_creal(float __x) {return __x;}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_creal(double __x) {return __x;}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_creal(long double __x) {return __x;}
+
+static float
+    _TG_ATTRS
+    __tg_creal(float _Complex __x) {return crealf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_creal(double _Complex __x) {return creal(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_creal(long double _Complex __x) {return creall(__x);}
+
+#undef creal
+#define creal(__x) __tg_creal(__tg_promote1((__x))(__x))
+
+#undef _TG_ATTRSp
+#undef _TG_ATTRS
+
+#endif /* __cplusplus */
+#endif /* __TGMATH_H */
diff --git a/clang/lib/Headers/tmmintrin.h b/clang/lib/Headers/tmmintrin.h
new file mode 100644
index 0000000..a62c6cc
--- /dev/null
+++ b/clang/lib/Headers/tmmintrin.h
@@ -0,0 +1,225 @@
+/*===---- tmmintrin.h - SSSE3 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __TMMINTRIN_H
+#define __TMMINTRIN_H
+
+#ifndef __SSSE3__
+#error "SSSE3 instruction set not enabled"
+#else
+
+#include <pmmintrin.h>
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi8(__m64 a)
+{
+    return (__m64)__builtin_ia32_pabsb((__v8qi)a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi8(__m128i a)
+{
+    return (__m128i)__builtin_ia32_pabsb128((__v16qi)a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi16(__m64 a)
+{
+    return (__m64)__builtin_ia32_pabsw((__v4hi)a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi16(__m128i a)
+{
+    return (__m128i)__builtin_ia32_pabsw128((__v8hi)a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi32(__m64 a)
+{
+    return (__m64)__builtin_ia32_pabsd((__v2si)a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi32(__m128i a)
+{
+    return (__m128i)__builtin_ia32_pabsd128((__v4si)a);
+}
+
+#define _mm_alignr_epi8(a, b, n) __extension__ ({ \
+  __m128i __a = (a); \
+  __m128i __b = (b); \
+  (__m128i)__builtin_ia32_palignr128((__v16qi)__a, (__v16qi)__b, (n)); })
+
+#define _mm_alignr_pi8(a, b, n) __extension__ ({ \
+  __m64 __a = (a); \
+  __m64 __b = (b); \
+  (__m64)__builtin_ia32_palignr((__v8qi)__a, (__v8qi)__b, (n)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phaddw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_epi32(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phaddd128((__v4si)a, (__v4si)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phaddw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pi32(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phaddd((__v2si)a, (__v2si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadds_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phaddsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadds_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phaddsw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phsubw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_epi32(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phsubd128((__v4si)a, (__v4si)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phsubw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pi32(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phsubd((__v2si)a, (__v2si)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsubs_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_phsubsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsubs_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_phsubsw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maddubs_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_pmaddubsw128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_maddubs_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_pmaddubsw((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhrs_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_pmulhrsw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhrs_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_pmulhrsw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shuffle_epi8(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_pshufb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_shuffle_pi8(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_pshufb((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi8(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_psignb128((__v16qi)a, (__v16qi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi16(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_psignw128((__v8hi)a, (__v8hi)b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi32(__m128i a, __m128i b)
+{
+    return (__m128i)__builtin_ia32_psignd128((__v4si)a, (__v4si)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi8(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_psignb((__v8qi)a, (__v8qi)b);  
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi16(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_psignw((__v4hi)a, (__v4hi)b);  
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi32(__m64 a, __m64 b)
+{
+    return (__m64)__builtin_ia32_psignd((__v2si)a, (__v2si)b);
+}
+
+#endif /* __SSSE3__ */
+
+#endif /* __TMMINTRIN_H */
diff --git a/clang/lib/Headers/unwind.h b/clang/lib/Headers/unwind.h
new file mode 100644
index 0000000..a065920
--- /dev/null
+++ b/clang/lib/Headers/unwind.h
@@ -0,0 +1,124 @@
+/*===---- unwind.h - Stack unwinding ----------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* See "Data Definitions for libgcc_s" in the Linux Standard Base.*/
+
+#if __has_include_next(<unwind.h>)
+/* Darwin and libunwind provide an unwind.h. If that's available, use
+ * it. libunwind wraps some of its definitions in #ifdef _GNU_SOURCE,
+ * so define that around the include.*/
+# ifndef _GNU_SOURCE
+#  define _SHOULD_UNDEFINE_GNU_SOURCE
+#  define _GNU_SOURCE
+# endif
+// libunwind's unwind.h reflects the current visibility.  However, Mozilla
+// builds with -fvisibility=hidden and relies on gcc's unwind.h to reset the
+// visibility to default and export its contents.  gcc also allows users to
+// override its override by #defining HIDE_EXPORTS (but note, this only obeys
+// the user's -fvisibility setting; it doesn't hide any exports on its own).  We
+// imitate gcc's header here:
+# ifdef HIDE_EXPORTS
+#  include_next <unwind.h>
+# else
+#  pragma GCC visibility push(default)
+#  include_next <unwind.h>
+#  pragma GCC visibility pop
+# endif
+# ifdef _SHOULD_UNDEFINE_GNU_SOURCE
+#  undef _GNU_SOURCE
+#  undef _SHOULD_UNDEFINE_GNU_SOURCE
+# endif
+#else
+
+#include <stdint.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* It is a bit strange for a header to play with the visibility of the
+   symbols it declares, but this matches gcc's behavior and some programs
+   depend on it */
+#pragma GCC visibility push(default)
+
+struct _Unwind_Context;
+typedef enum {
+  _URC_NO_REASON = 0,
+  _URC_FOREIGN_EXCEPTION_CAUGHT = 1,
+
+  _URC_FATAL_PHASE2_ERROR = 2,
+  _URC_FATAL_PHASE1_ERROR = 3,
+  _URC_NORMAL_STOP = 4,
+
+  _URC_END_OF_STACK = 5,
+  _URC_HANDLER_FOUND = 6,
+  _URC_INSTALL_CONTEXT = 7,
+  _URC_CONTINUE_UNWIND = 8
+} _Unwind_Reason_Code;
+
+
+#ifdef __arm__
+
+typedef enum { 
+  _UVRSC_CORE = 0,        /* integer register */ 
+  _UVRSC_VFP = 1,         /* vfp */ 
+  _UVRSC_WMMXD = 3,       /* Intel WMMX data register */ 
+  _UVRSC_WMMXC = 4        /* Intel WMMX control register */ 
+} _Unwind_VRS_RegClass; 
+
+typedef enum { 
+  _UVRSD_UINT32 = 0,  
+  _UVRSD_VFPX = 1,  
+  _UVRSD_UINT64 = 3,  
+  _UVRSD_FLOAT = 4,  
+  _UVRSD_DOUBLE = 5 
+} _Unwind_VRS_DataRepresentation; 
+
+typedef enum { 
+  _UVRSR_OK = 0,  
+  _UVRSR_NOT_IMPLEMENTED = 1,  
+  _UVRSR_FAILED = 2  
+} _Unwind_VRS_Result; 
+
+_Unwind_VRS_Result _Unwind_VRS_Get(_Unwind_Context *context,
+  _Unwind_VRS_RegClass regclass,
+  uint32_t regno,
+  _Unwind_VRS_DataRepresentation representation,
+  void *valuep);
+
+#else
+
+uintptr_t _Unwind_GetIP(struct _Unwind_Context* context);
+
+#endif
+
+typedef _Unwind_Reason_Code (*_Unwind_Trace_Fn)(struct _Unwind_Context*, void*);
+_Unwind_Reason_Code _Unwind_Backtrace(_Unwind_Trace_Fn, void*);
+
+#pragma GCC visibility pop
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
diff --git a/clang/lib/Headers/varargs.h b/clang/lib/Headers/varargs.h
new file mode 100644
index 0000000..b5477d0
--- /dev/null
+++ b/clang/lib/Headers/varargs.h
@@ -0,0 +1,26 @@
+/*===---- varargs.h - Variable argument handling -------------------------------------===
+*
+* Permission is hereby granted, free of charge, to any person obtaining a copy
+* of this software and associated documentation files (the "Software"), to deal
+* in the Software without restriction, including without limitation the rights
+* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+* copies of the Software, and to permit persons to whom the Software is
+* furnished to do so, subject to the following conditions:
+*
+* The above copyright notice and this permission notice shall be included in
+* all copies or substantial portions of the Software.
+*
+* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+* THE SOFTWARE.
+*
+*===-----------------------------------------------------------------------===
+*/
+#ifndef __VARARGS_H
+#define __VARARGS_H
+  #error "Please use <stdarg.h> instead of <varargs.h>"
+#endif
diff --git a/clang/lib/Headers/wmmintrin.h b/clang/lib/Headers/wmmintrin.h
new file mode 100644
index 0000000..8f58850
--- /dev/null
+++ b/clang/lib/Headers/wmmintrin.h
@@ -0,0 +1,67 @@
+/*===---- wmmintrin.h - AES intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _WMMINTRIN_H
+#define _WMMINTRIN_H
+
+#if !defined (__AES__)
+# error "AES instructions not enabled"
+#else
+
+#include <xmmintrin.h>
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesenc_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesenc128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesenclast_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesenclast128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesdec_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesdec128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesdeclast_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesdeclast128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesimc_si128(__m128i __V)
+{
+  return (__m128i)__builtin_ia32_aesimc128(__V);
+}
+
+#define _mm_aeskeygenassist_si128(C, R) \
+  __builtin_ia32_aeskeygenassist128((C), (R))
+
+#endif /* __AES__ */
+#endif /* _WMMINTRIN_H */
diff --git a/clang/lib/Headers/x86intrin.h b/clang/lib/Headers/x86intrin.h
new file mode 100644
index 0000000..f5e4d88
--- /dev/null
+++ b/clang/lib/Headers/x86intrin.h
@@ -0,0 +1,55 @@
+/*===---- x86intrin.h - X86 intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#define __X86INTRIN_H
+
+#include <immintrin.h>
+
+#ifdef __3dNOW__
+#include <mm3dnow.h>
+#endif
+
+#ifdef __BMI__
+#include <bmiintrin.h>
+#endif
+
+#ifdef __BMI2__
+#include <bmi2intrin.h>
+#endif
+
+#ifdef __LZCNT__
+#include <lzcntintrin.h>
+#endif
+
+#ifdef __POPCNT__
+#include <popcntintrin.h>
+#endif
+
+#ifdef __FMA4__
+#include <fma4intrin.h>
+#endif
+
+// FIXME: SSE4A, XOP, LWP, ABM
+
+#endif /* __X86INTRIN_H */
diff --git a/clang/lib/Headers/xmmintrin.h b/clang/lib/Headers/xmmintrin.h
new file mode 100644
index 0000000..e616157
--- /dev/null
+++ b/clang/lib/Headers/xmmintrin.h
@@ -0,0 +1,990 @@
+/*===---- xmmintrin.h - SSE intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __XMMINTRIN_H
+#define __XMMINTRIN_H
+ 
+#ifndef __SSE__
+#error "SSE instruction set not enabled"
+#else
+
+#include <mmintrin.h>
+
+typedef int __v4si __attribute__((__vector_size__(16)));
+typedef float __v4sf __attribute__((__vector_size__(16)));
+typedef float __m128 __attribute__((__vector_size__(16)));
+
+// This header should only be included in a hosted environment as it depends on
+// a standard library to provide allocation routines.
+#if __STDC_HOSTED__
+#include <mm_malloc.h>
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_add_ss(__m128 a, __m128 b)
+{
+  a[0] += b[0];
+  return a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_add_ps(__m128 a, __m128 b)
+{
+  return a + b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_ss(__m128 a, __m128 b)
+{
+  a[0] -= b[0];
+  return a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_ps(__m128 a, __m128 b)
+{
+  return a - b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_ss(__m128 a, __m128 b)
+{
+  a[0] *= b[0];
+  return a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_ps(__m128 a, __m128 b)
+{
+  return a * b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_div_ss(__m128 a, __m128 b)
+{
+  a[0] /= b[0];
+  return a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_div_ps(__m128 a, __m128 b)
+{
+  return a / b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_ss(__m128 a)
+{
+  return __builtin_ia32_sqrtss(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_ps(__m128 a)
+{
+  return __builtin_ia32_sqrtps(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp_ss(__m128 a)
+{
+  return __builtin_ia32_rcpss(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp_ps(__m128 a)
+{
+  return __builtin_ia32_rcpps(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt_ss(__m128 a)
+{
+  return __builtin_ia32_rsqrtss(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt_ps(__m128 a)
+{
+  return __builtin_ia32_rsqrtps(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_min_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_minss(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_min_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_minps(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_max_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_maxss(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_max_ps(__m128 a, __m128 b)
+{
+  return __builtin_ia32_maxps(a, b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_and_ps(__m128 a, __m128 b)
+{
+  return (__m128)((__v4si)a & (__v4si)b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_ps(__m128 a, __m128 b)
+{
+  return (__m128)(~(__v4si)a & (__v4si)b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_or_ps(__m128 a, __m128 b)
+{
+  return (__m128)((__v4si)a | (__v4si)b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_xor_ps(__m128 a, __m128 b)
+{
+  return (__m128)((__v4si)a ^ (__v4si)b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(b, a, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(b, a, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(b, a, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(b, a, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 4);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 4);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(b, a, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(b, a, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(b, a, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(b, a, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_ss(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpss(a, b, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_ps(__m128 a, __m128 b)
+{
+  return (__m128)__builtin_ia32_cmpps(a, b, 3);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comieq_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comieq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comilt_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comilt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comile_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comile(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comigt_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comigt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comige_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comige(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comineq_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_comineq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomieq_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomieq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomilt_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomilt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomile_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomile(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomigt_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomigt(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomige_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomige(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomineq_ss(__m128 a, __m128 b)
+{
+  return __builtin_ia32_ucomineq(a, b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_si32(__m128 a)
+{
+  return __builtin_ia32_cvtss2si(a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_ss2si(__m128 a)
+{
+  return _mm_cvtss_si32(a);
+}
+
+#ifdef __x86_64__
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_si64(__m128 a)
+{
+  return __builtin_ia32_cvtss2si64(a);
+}
+
+#endif
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi32(__m128 a)
+{
+  return (__m64)__builtin_ia32_cvtps2pi(a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_ps2pi(__m128 a)
+{
+  return _mm_cvtps_pi32(a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvttss_si32(__m128 a)
+{
+  return a[0];
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtt_ss2si(__m128 a)
+{
+  return _mm_cvttss_si32(a);
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvttss_si64(__m128 a)
+{
+  return a[0];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvttps_pi32(__m128 a)
+{
+  return (__m64)__builtin_ia32_cvttps2pi(a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtt_ps2pi(__m128 a)
+{
+  return _mm_cvttps_pi32(a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_ss(__m128 a, int b)
+{
+  a[0] = b;
+  return a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_si2ss(__m128 a, int b)
+{
+  return _mm_cvtsi32_ss(a, b);
+}
+
+#ifdef __x86_64__
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_ss(__m128 a, long long b)
+{
+  a[0] = b;
+  return a;
+}
+
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32_ps(__m128 a, __m64 b)
+{
+  return __builtin_ia32_cvtpi2ps(a, (__v2si)b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_pi2ps(__m128 a, __m64 b)
+{
+  return _mm_cvtpi32_ps(a, b);
+}
+
+static __inline__ float __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_f32(__m128 a)
+{
+  return a[0];
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadh_pi(__m128 a, const __m64 *p)
+{
+  typedef float __mm_loadh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_loadh_pi_struct {
+    __mm_loadh_pi_v2f32 u;
+  } __attribute__((__packed__, __may_alias__));
+  __mm_loadh_pi_v2f32 b = ((struct __mm_loadh_pi_struct*)p)->u;
+  __m128 bb = __builtin_shufflevector(b, b, 0, 1, 0, 1);
+  return __builtin_shufflevector(a, bb, 0, 1, 4, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_pi(__m128 a, const __m64 *p)
+{
+  typedef float __mm_loadl_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_loadl_pi_struct {
+    __mm_loadl_pi_v2f32 u;
+  } __attribute__((__packed__, __may_alias__));
+  __mm_loadl_pi_v2f32 b = ((struct __mm_loadl_pi_struct*)p)->u;
+  __m128 bb = __builtin_shufflevector(b, b, 0, 1, 0, 1);
+  return __builtin_shufflevector(a, bb, 4, 5, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load_ss(const float *p)
+{
+  struct __mm_load_ss_struct {
+    float u;
+  } __attribute__((__packed__, __may_alias__));
+  float u = ((struct __mm_load_ss_struct*)p)->u;
+  return (__m128){ u, 0, 0, 0 };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load1_ps(const float *p)
+{
+  struct __mm_load1_ps_struct {
+    float u;
+  } __attribute__((__packed__, __may_alias__));
+  float u = ((struct __mm_load1_ps_struct*)p)->u;
+  return (__m128){ u, u, u, u };
+}
+
+#define        _mm_load_ps1(p) _mm_load1_ps(p)
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load_ps(const float *p)
+{
+  return *(__m128*)p;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_ps(const float *p)
+{
+  struct __loadu_ps {
+    __m128 v;
+  } __attribute__((__packed__, __may_alias__));
+  return ((struct __loadu_ps*)p)->v;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadr_ps(const float *p)
+{
+  __m128 a = _mm_load_ps(p);
+  return __builtin_shufflevector(a, a, 3, 2, 1, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ss(float w)
+{
+  return (__m128){ w, 0, 0, 0 };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_ps(float w)
+{
+  return (__m128){ w, w, w, w };
+}
+
+// Microsoft specific.
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ps1(float w)
+{
+    return _mm_set1_ps(w);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ps(float z, float y, float x, float w)
+{
+  return (__m128){ w, x, y, z };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_ps(float z, float y, float x, float w)
+{
+  return (__m128){ z, y, x, w };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__))
+_mm_setzero_ps(void)
+{
+  return (__m128){ 0, 0, 0, 0 };
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_storeh_pi(__m64 *p, __m128 a)
+{
+  __builtin_ia32_storehps((__v2si *)p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_storel_pi(__m64 *p, __m128 a)
+{
+  __builtin_ia32_storelps((__v2si *)p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_store_ss(float *p, __m128 a)
+{
+  struct __mm_store_ss_struct {
+    float u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store_ss_struct*)p)->u = a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_ps(float *p, __m128 a)
+{
+  __builtin_ia32_storeups(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store1_ps(float *p, __m128 a)
+{
+  a = __builtin_shufflevector(a, a, 0, 0, 0, 0);
+  _mm_storeu_ps(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_ps1(float *p, __m128 a)
+{
+    return _mm_store1_ps(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_ps(float *p, __m128 a)
+{
+  *(__m128 *)p = a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storer_ps(float *p, __m128 a)
+{
+  a = __builtin_shufflevector(a, a, 3, 2, 1, 0);
+  _mm_store_ps(p, a);
+}
+
+#define _MM_HINT_T0 3
+#define _MM_HINT_T1 2
+#define _MM_HINT_T2 1
+#define _MM_HINT_NTA 0
+
+/* FIXME: We have to #define this because "sel" must be a constant integer, and
+   Sema doesn't do any form of constant propagation yet. */
+
+#define _mm_prefetch(a, sel) (__builtin_prefetch((void *)(a), 0, (sel)))
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_pi(__m64 *p, __m64 a)
+{
+  __builtin_ia32_movntq(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_ps(float *p, __m128 a)
+{
+  __builtin_ia32_movntps(p, a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_sfence(void)
+{
+  __builtin_ia32_sfence();
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_extract_pi16(__m64 a, int n)
+{
+  __v4hi b = (__v4hi)a;
+  return (unsigned short)b[n & 3];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_insert_pi16(__m64 a, int d, int n)
+{
+   __v4hi b = (__v4hi)a;
+   b[n & 3] = d;
+   return (__m64)b;
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_max_pi16(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pmaxsw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_max_pu8(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pmaxub((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_min_pi16(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pminsw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_min_pu8(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pminub((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_pi8(__m64 a)
+{
+  return __builtin_ia32_pmovmskb((__v8qi)a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_pu16(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pmulhuw((__v4hi)a, (__v4hi)b);  
+}
+
+#define _mm_shuffle_pi16(a, n) __extension__ ({ \
+  __m64 __a = (a); \
+  (__m64)__builtin_ia32_pshufw((__v4hi)__a, (n)); })
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskmove_si64(__m64 d, __m64 n, char *p)
+{
+  __builtin_ia32_maskmovq((__v8qi)d, (__v8qi)n, p);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_avg_pu8(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pavgb((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_avg_pu16(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_pavgw((__v4hi)a, (__v4hi)b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sad_pu8(__m64 a, __m64 b)
+{
+  return (__m64)__builtin_ia32_psadbw((__v8qi)a, (__v8qi)b);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_getcsr(void)
+{
+  return __builtin_ia32_stmxcsr();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_setcsr(unsigned int i)
+{
+  __builtin_ia32_ldmxcsr(i);
+}
+
+#define _mm_shuffle_ps(a, b, mask) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_shufflevector((__v4sf)__a, (__v4sf)__b, \
+                                  (mask) & 0x3, ((mask) & 0xc) >> 2, \
+                                  (((mask) & 0x30) >> 4) + 4, \
+                                  (((mask) & 0xc0) >> 6) + 4); })
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_ps(__m128 a, __m128 b)
+{
+  return __builtin_shufflevector(a, b, 2, 6, 3, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_ps(__m128 a, __m128 b)
+{
+  return __builtin_shufflevector(a, b, 0, 4, 1, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_move_ss(__m128 a, __m128 b)
+{
+  return __builtin_shufflevector(a, b, 4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movehl_ps(__m128 a, __m128 b)
+{
+  return __builtin_shufflevector(a, b, 6, 7, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movelh_ps(__m128 a, __m128 b)
+{
+  return __builtin_shufflevector(a, b, 0, 1, 4, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi16_ps(__m64 a)
+{
+  __m64 b, c;
+  __m128 r;
+
+  b = _mm_setzero_si64();
+  b = _mm_cmpgt_pi16(b, a);
+  c = _mm_unpackhi_pi16(a, b);  
+  r = _mm_setzero_ps();
+  r = _mm_cvtpi32_ps(r, c);
+  r = _mm_movelh_ps(r, r);
+  c = _mm_unpacklo_pi16(a, b);  
+  r = _mm_cvtpi32_ps(r, c);
+
+  return r;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpu16_ps(__m64 a)
+{
+  __m64 b, c;
+  __m128 r;
+
+  b = _mm_setzero_si64();
+  c = _mm_unpackhi_pi16(a, b);  
+  r = _mm_setzero_ps();
+  r = _mm_cvtpi32_ps(r, c);
+  r = _mm_movelh_ps(r, r);
+  c = _mm_unpacklo_pi16(a, b);  
+  r = _mm_cvtpi32_ps(r, c);
+
+  return r;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi8_ps(__m64 a)
+{
+  __m64 b;
+  
+  b = _mm_setzero_si64();
+  b = _mm_cmpgt_pi8(b, a);
+  b = _mm_unpacklo_pi8(a, b);
+
+  return _mm_cvtpi16_ps(b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpu8_ps(__m64 a)
+{
+  __m64 b;
+  
+  b = _mm_setzero_si64();
+  b = _mm_unpacklo_pi8(a, b);
+
+  return _mm_cvtpi16_ps(b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32x2_ps(__m64 a, __m64 b)
+{
+  __m128 c;
+  
+  c = _mm_setzero_ps();  
+  c = _mm_cvtpi32_ps(c, b);
+  c = _mm_movelh_ps(c, c);
+
+  return _mm_cvtpi32_ps(c, a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi16(__m128 a)
+{
+  __m64 b, c;
+  
+  b = _mm_cvtps_pi32(a);
+  a = _mm_movehl_ps(a, a);
+  c = _mm_cvtps_pi32(a);
+  
+  return _mm_packs_pi16(b, c);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi8(__m128 a)
+{
+  __m64 b, c;
+  
+  b = _mm_cvtps_pi16(a);
+  c = _mm_setzero_si64();
+  
+  return _mm_packs_pi16(b, c);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_ps(__m128 a)
+{
+  return __builtin_ia32_movmskps(a);
+}
+
+#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))
+
+#define _MM_EXCEPT_INVALID    (0x0001)
+#define _MM_EXCEPT_DENORM     (0x0002)
+#define _MM_EXCEPT_DIV_ZERO   (0x0004)
+#define _MM_EXCEPT_OVERFLOW   (0x0008)
+#define _MM_EXCEPT_UNDERFLOW  (0x0010)
+#define _MM_EXCEPT_INEXACT    (0x0020)
+#define _MM_EXCEPT_MASK       (0x003f)
+
+#define _MM_MASK_INVALID      (0x0080)
+#define _MM_MASK_DENORM       (0x0100)
+#define _MM_MASK_DIV_ZERO     (0x0200)
+#define _MM_MASK_OVERFLOW     (0x0400)
+#define _MM_MASK_UNDERFLOW    (0x0800)
+#define _MM_MASK_INEXACT      (0x1000)
+#define _MM_MASK_MASK         (0x1f80)
+
+#define _MM_ROUND_NEAREST     (0x0000)
+#define _MM_ROUND_DOWN        (0x2000)
+#define _MM_ROUND_UP          (0x4000)
+#define _MM_ROUND_TOWARD_ZERO (0x6000)
+#define _MM_ROUND_MASK        (0x6000)
+
+#define _MM_FLUSH_ZERO_MASK   (0x8000)
+#define _MM_FLUSH_ZERO_ON     (0x8000)
+#define _MM_FLUSH_ZERO_OFF    (0x0000)
+
+#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
+#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
+#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
+#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)
+
+#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
+#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
+#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
+#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))
+
+#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
+do { \
+  __m128 tmp3, tmp2, tmp1, tmp0; \
+  tmp0 = _mm_unpacklo_ps((row0), (row1)); \
+  tmp2 = _mm_unpacklo_ps((row2), (row3)); \
+  tmp1 = _mm_unpackhi_ps((row0), (row1)); \
+  tmp3 = _mm_unpackhi_ps((row2), (row3)); \
+  (row0) = _mm_movelh_ps(tmp0, tmp2); \
+  (row1) = _mm_movehl_ps(tmp2, tmp0); \
+  (row2) = _mm_movelh_ps(tmp1, tmp3); \
+  (row3) = _mm_movehl_ps(tmp3, tmp1); \
+} while (0)
+
+/* Aliases for compatibility. */
+#define _m_pextrw _mm_extract_pi16
+#define _m_pinsrw _mm_insert_pi16
+#define _m_pmaxsw _mm_max_pi16
+#define _m_pmaxub _mm_max_pu8
+#define _m_pminsw _mm_min_pi16
+#define _m_pminub _mm_min_pu8
+#define _m_pmovmskb _mm_movemask_pi8
+#define _m_pmulhuw _mm_mulhi_pu16
+#define _m_pshufw _mm_shuffle_pi16
+#define _m_maskmovq _mm_maskmove_si64
+#define _m_pavgb _mm_avg_pu8
+#define _m_pavgw _mm_avg_pu16
+#define _m_psadbw _mm_sad_pu8
+#define _m_ _mm_
+#define _m_ _mm_
+
+/* Ugly hack for backwards-compatibility (compatible with gcc) */
+#ifdef __SSE2__
+#include <emmintrin.h>
+#endif
+
+#endif /* __SSE__ */
+
+#endif /* __XMMINTRIN_H */