2006-10-22 H.J. Lu <hongjiu.lu@intel.com>

	* config.gcc (i[34567]86-*-*): Add tmmintrin.h to extra_headers.
	(x86_64-*-*): Likewise.

	* config/i386/i386.c (pta_flags): Add PTA_SSSE3.
	(override_options): Check SSSE3.
	(ix86_builtins): Add IX86_BUILTIN_PHADDW, IX86_BUILTIN_PHADDD,
	IX86_BUILTIN_PHADDSW, IX86_BUILTIN_PHSUBW, IX86_BUILTIN_PHSUBD,
	IX86_BUILTIN_PHSUBSW, IX86_BUILTIN_PMADDUBSW,
	IX86_BUILTIN_PMULHRSW, IX86_BUILTIN_PSHUFB,
	IX86_BUILTIN_PSIGNB, IX86_BUILTIN_PSIGNW, IX86_BUILTIN_PSIGND,
	IX86_BUILTIN_PALIGNR, IX86_BUILTIN_PABSB, IX86_BUILTIN_PABSW,
	IX86_BUILTIN_PABSD, IX86_BUILTIN_PHADDW128,
	IX86_BUILTIN_PHADDD128, IX86_BUILTIN_PHADDSW128,
	IX86_BUILTIN_PHSUBW128, IX86_BUILTIN_PHSUBD128,
	IX86_BUILTIN_PHSUBSW128, IX86_BUILTIN_PMADDUBSW128,
	IX86_BUILTIN_PMULHRSW128, IX86_BUILTIN_PSHUFB128,
	IX86_BUILTIN_PSIGNB128, IX86_BUILTIN_PSIGNW128,
	IX86_BUILTIN_PSIGND128, IX86_BUILTIN_PALIGNR128,
	IX86_BUILTIN_PABSB128, IX86_BUILTIN_PABSW128 and
	IX86_BUILTIN_PABSD128.
	(bdesc_2arg): Add SSSE3.
	(bdesc_1arg): Likewise.
	(ix86_init_mmx_sse_builtins): Support SSSE3.
	(ix86_expand_builtin): Likewise.
	* config/i386/i386.h (TARGET_CPU_CPP_BUILTINS): Likewise.

	* config/i386/i386.md (UNSPEC_PSHUFB): New.
	(UNSPEC_PSIGN): Likewise.
	(UNSPEC_PALIGNR): Likewise.
	Include mmx.md before sse.md.

	* config/i386/i386.opt: Add -mssse3.

	* config/i386/sse.md (ssse3_phaddwv8hi3): New pattern for SSSE3.
	(ssse3_phaddwv4hi3): Likewise.
	(ssse3_phadddv4si3): Likewise.
	(ssse3_phadddv2si3): Likewise.
	(ssse3_phaddswv8hi3): Likewise.
	(ssse3_phaddswv4hi3): Likewise.
	(ssse3_phsubwv8hi3): Likewise.
	(ssse3_phsubwv4hi3): Likewise.
	(ssse3_phsubdv4si3): Likewise.
	(ssse3_phsubdv2si3): Likewise.
	(ssse3_phsubswv8hi3): Likewise.
	(ssse3_phsubswv4hi3): Likewise.
	(ssse3_pmaddubswv8hi3): Likewise.
	(ssse3_pmaddubswv4hi3): Likewise.
	(ssse3_pmulhrswv8hi3): Likewise.
	(ssse3_pmulhrswv4hi3): Likewise.
	(ssse3_pshufbv16qi3): Likewise.
	(ssse3_pshufbv8qi3): Likewise.
	(ssse3_psign<mode>3): Likewise.
	(ssse3_psign<mode>3): Likewise.
	(ssse3_palignrti): Likewise.
	(ssse3_palignrdi): Likewise.
	(abs<mode>2): Likewise.
	(abs<mode>2): Likewise.

	* config/i386/tmmintrin.h: New file.

	* doc/extend.texi: Document SSSE3 built-in functions.

	* doc/invoke.texi: Document -mssse3/-mno-ssse3 switches.


git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@117958 138bc75d-0d04-0410-961f-82ee72b054a4

1 files changed, 448 insertions, 0 deletions

diff --git a/gcc/config/i386/tmmintrin.h b/gcc/config/i386/tmmintrin.h
new file mode 100644
index 00000000000..e1bedc56135
--- /dev/null
+++ b/gcc/config/i386/tmmintrin.h
@@ -0,0 +1,448 @@
+/* Copyright (C) 2006 Free Software Foundation, Inc.
+
+   This file is part of GCC.
+
+   GCC is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2, or (at your option)
+   any later version.
+
+   GCC is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with GCC; see the file COPYING.  If not, write to
+   the Free Software Foundation, 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you include this header file into source
+   files compiled by GCC, this header file does not by itself cause
+   the resulting executable to be covered by the GNU General Public
+   License.  This exception does not however invalidate any other
+   reasons why the executable file might be covered by the GNU General
+   Public License.  */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.1.  */
+
+#ifndef _TMMINTRIN_H_INCLUDED
+#define _TMMINTRIN_H_INCLUDED
+
+#ifdef __SSSE3__
+#include <pmmintrin.h>
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadd_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadd_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadds_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadd_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadd_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddd ((__v2si)__X, (__v2si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadds_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsub_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsub_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsubs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsub_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsub_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubd ((__v2si)__X, (__v2si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsubs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_maddubs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pmaddubsw128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_maddubs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pmaddubsw ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_mulhrs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pmulhrsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_mulhrs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pmulhrsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_shuffle_epi8 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pshufb128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_shuffle_pi8 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pshufb ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi8 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignb128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi8 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignb ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignd ((__v2si)__X, (__v2si)__Y);
+}
+
+#define _mm_alignr_epi8(__X, __Y, __N) \
+  ((__m128i)__builtin_ia32_palignr128 ((__v2di) __X, (__v2di) __Y, (__N) * 8))
+
+#define _mm_alignr_pi8(__X, __Y, __N) \
+  ((__m64)__builtin_ia32_palignr ((long long) (__X), (long long) (__Y), (__N) * 8))
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi8 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsb128 ((__v16qi)__X);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi16 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsw128 ((__v8hi)__X);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi32 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsd128 ((__v4si)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi8 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsb ((__v8qi)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi16 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsw ((__v4hi)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi32 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsd ((__v2si)__X);
+}
+
+#endif /* __SSSE3__ */
+
+#endif /* _TMMINTRIN_H_INCLUDED */
+/* Copyright (C) 2006 Free Software Foundation, Inc.
+
+   This file is part of GCC.
+
+   GCC is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2, or (at your option)
+   any later version.
+
+   GCC is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with GCC; see the file COPYING.  If not, write to
+   the Free Software Foundation, 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you include this header file into source
+   files compiled by GCC, this header file does not by itself cause
+   the resulting executable to be covered by the GNU General Public
+   License.  This exception does not however invalidate any other
+   reasons why the executable file might be covered by the GNU General
+   Public License.  */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.1.  */
+
+#ifndef _TMMINTRIN_H_INCLUDED
+#define _TMMINTRIN_H_INCLUDED
+
+#ifdef __SSSE3__
+#include <pmmintrin.h>
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadd_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadd_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hadds_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phaddsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadd_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadd_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddd ((__v2si)__X, (__v2si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hadds_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phaddsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsub_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsub_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_hsubs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_phsubsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsub_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsub_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubd ((__v2si)__X, (__v2si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_hsubs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_phsubsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_maddubs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pmaddubsw128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_maddubs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pmaddubsw ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_mulhrs_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pmulhrsw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_mulhrs_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pmulhrsw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_shuffle_epi8 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_pshufb128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_shuffle_pi8 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_pshufb ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi8 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignb128 ((__v16qi)__X, (__v16qi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi16 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignw128 ((__v8hi)__X, (__v8hi)__Y);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_sign_epi32 (__m128i __X, __m128i __Y)
+{
+  return (__m128i) __builtin_ia32_psignd128 ((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi8 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignb ((__v8qi)__X, (__v8qi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi16 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignw ((__v4hi)__X, (__v4hi)__Y);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_sign_pi32 (__m64 __X, __m64 __Y)
+{
+  return (__m64) __builtin_ia32_psignd ((__v2si)__X, (__v2si)__Y);
+}
+
+#define _mm_alignr_epi8(__X, __Y, __N) \
+  ((__m128i)__builtin_ia32_palignr128 ((__v2di) __X, (__v2di) __Y, (__N) * 8))
+
+#define _mm_alignr_pi8(__X, __Y, __N) \
+  ((__m64)__builtin_ia32_palignr ((long long) (__X), (long long) (__Y), (__N) * 8))
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi8 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsb128 ((__v16qi)__X);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi16 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsw128 ((__v8hi)__X);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_abs_epi32 (__m128i __X)
+{
+  return (__m128i) __builtin_ia32_pabsd128 ((__v4si)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi8 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsb ((__v8qi)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi16 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsw ((__v4hi)__X);
+}
+
+static __inline __m64 __attribute__((__always_inline__))
+_mm_abs_pi32 (__m64 __X)
+{
+  return (__m64) __builtin_ia32_pabsd ((__v2si)__X);
+}
+
+#endif /* __SSSE3__ */
+
+#endif /* _TMMINTRIN_H_INCLUDED */