+2014-11-21 Ilya Tocar <ilya.tocar@intel.com>
+
+ * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512IFMA_SET,
+ OPTION_MASK_ISA_AVX512IFMA_UNSET): New.
+ (ix86_handle_option): Handle OPT_mavx512ifma.
+ * config.gcc: Add avx512ifmaintrin.h, avx512ifmavlintrin.h.
+ * config/i386/avx512ifmaintrin.h: New file.
+ * config/i386/avx512ifmaivlntrin.h: Ditto.
+ * config/i386/cpuid.h (bit_AVX512IFMA): New.
+ * config/i386/driver-i386.c (host_detect_local_cpu): Detect
+ avx512ifma.
+ * config/i386/i386-c.c (ix86_target_macros_internal): Define
+ __AVX512IFMA__.
+ * config/i386/i386.c (ix86_target_string): Add -mavx512ifma.
+ (PTA_AVX512IFMA): Define.
+ (ix86_option_override_internal): Handle new options.
+ (ix86_valid_target_attribute_inner_p): Add avx512ifma.
+ (ix86_builtins): Add IX86_BUILTIN_VPMADD52LUQ512,
+ IX86_BUILTIN_VPMADD52HUQ512, IX86_BUILTIN_VPMADD52LUQ256,
+ IX86_BUILTIN_VPMADD52HUQ256, IX86_BUILTIN_VPMADD52LUQ128,
+ IX86_BUILTIN_VPMADD52HUQ128, IX86_BUILTIN_VPMADD52LUQ512_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ512_MASKZ, IX86_BUILTIN_VPMADD52LUQ256_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ256_MASKZ, IX86_BUILTIN_VPMADD52LUQ128_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ128_MASKZ.
+ (bdesc_special_args): Add __builtin_ia32_vpmadd52luq512_mask,
+ __builtin_ia32_vpmadd52luq512_maskz,
+ __builtin_ia32_vpmadd52huq512_mask,
+ __builtin_ia32_vpmadd52huq512_maskx,
+ __builtin_ia32_vpmadd52luq256_mask,
+ __builtin_ia32_vpmadd52luq256_maskz,
+ __builtin_ia32_vpmadd52huq256_mask,
+ __builtin_ia32_vpmadd52huq256_maskz,
+ __builtin_ia32_vpmadd52luq128_mask,
+ __builtin_ia32_vpmadd52luq128_maskz,
+ __builtin_ia32_vpmadd52huq128_mask,
+ __builtin_ia32_vpmadd52huq128_maskz,
+ * config/i386/i386.h (TARGET_AVX512IFMA, TARGET_AVX512IFMA_P): Define.
+ * config/i386/i386.opt: Add mavx512ifma.
+ * config/i386/immintrin.h: Include avx512ifmaintrin.h,
+ avx512ifmavlintrin.h.
+ * config/i386/sse.md (unspec): Add UNSPEC_VPMADD52LUQ,
+ UNSPEC_VPMADD52HUQ.
+ (VPMADD52): New iterator.
+ (vpmadd52type): New attribute.
+ (vpamdd52huq<mode>_maskz): New.
+ (vpamdd52luq<mode>_maskz): Ditto.
+ (vpamdd52<vpmadd52type><mode><sd_maskz_name>): Ditto.
+ (vpamdd52<vpmadd52type><mode>_mask): Ditto.
+
2014-11-21 Alan Lawrence <alan.lawrence@arm.com>
Revert:
(OPTION_MASK_ISA_AVX512BW | OPTION_MASK_ISA_AVX512F_SET)
#define OPTION_MASK_ISA_AVX512VL_SET \
(OPTION_MASK_ISA_AVX512VL | OPTION_MASK_ISA_AVX512F_SET)
+#define OPTION_MASK_ISA_AVX512IFMA_SET \
+ (OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512F_SET)
#define OPTION_MASK_ISA_RTM_SET OPTION_MASK_ISA_RTM
#define OPTION_MASK_ISA_PRFCHW_SET OPTION_MASK_ISA_PRFCHW
#define OPTION_MASK_ISA_RDSEED_SET OPTION_MASK_ISA_RDSEED
#define OPTION_MASK_ISA_AVX512DQ_UNSET OPTION_MASK_ISA_AVX512DQ
#define OPTION_MASK_ISA_AVX512BW_UNSET OPTION_MASK_ISA_AVX512BW
#define OPTION_MASK_ISA_AVX512VL_UNSET OPTION_MASK_ISA_AVX512VL
+#define OPTION_MASK_ISA_AVX512IFMA_UNSET OPTION_MASK_ISA_AVX512IFMA
#define OPTION_MASK_ISA_RTM_UNSET OPTION_MASK_ISA_RTM
#define OPTION_MASK_ISA_PRFCHW_UNSET OPTION_MASK_ISA_PRFCHW
#define OPTION_MASK_ISA_RDSEED_UNSET OPTION_MASK_ISA_RDSEED
}
return true;
+ case OPT_mavx512ifma:
+ if (value)
+ {
+ opts->x_ix86_isa_flags |= OPTION_MASK_ISA_AVX512IFMA_SET;
+ opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_AVX512IFMA_SET;
+ }
+ else
+ {
+ opts->x_ix86_isa_flags &= ~OPTION_MASK_ISA_AVX512IFMA_UNSET;
+ opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_AVX512IFMA_UNSET;
+ }
+ return true;
+
case OPT_mfma:
if (value)
{
avx512cdintrin.h avx512erintrin.h avx512pfintrin.h
shaintrin.h clflushoptintrin.h xsavecintrin.h
xsavesintrin.h avx512dqintrin.h avx512bwintrin.h
- avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h"
+ avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h
+ avx512ifmaintrin.h avx512ifmavlintrin.h"
;;
x86_64-*-*)
cpu_type=i386
avx512cdintrin.h avx512erintrin.h avx512pfintrin.h
shaintrin.h clflushoptintrin.h xsavecintrin.h
xsavesintrin.h avx512dqintrin.h avx512bwintrin.h
- avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h"
+ avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h
+ avx512ifmaintrin.h avx512ifmavlintrin.h"
;;
ia64-*-*)
extra_headers=ia64intrin.h
--- /dev/null
+/* Copyright (C) 2013-2014 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 3, 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.
+
+ Under Section 7 of GPL version 3, you are granted additional
+ permissions described in the GCC Runtime Library Exception, version
+ 3.1, as published by the Free Software Foundation.
+
+ You should have received a copy of the GNU General Public License and
+ a copy of the GCC Runtime Library Exception along with this program;
+ see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef _IMMINTRIN_H_INCLUDED
+#error "Never use <avx512ifmaintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef _AVX512IFMAINTRIN_H_INCLUDED
+#define _AVX512IFMAINTRIN_H_INCLUDED
+
+#ifndef __AVX512IFMA__
+#pragma GCC push_options
+#pragma GCC target("avx512ifma")
+#define __DISABLE_AVX512IFMA__
+#endif /* __AVX512IFMA__ */
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_madd52lo_epu64 (__m512i __X, __m512i __Y, __m512i __Z)
+{
+ return (__m512i) __builtin_ia32_vpmadd52luq512_mask ((__v8di) __X,
+ (__v8di) __Y,
+ (__v8di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_madd52hi_epu64 (__m512i __X, __m512i __Y, __m512i __Z)
+{
+ return (__m512i) __builtin_ia32_vpmadd52huq512_mask ((__v8di) __X,
+ (__v8di) __Y,
+ (__v8di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask_madd52lo_epu64 (__m512i __W, __mmask8 __M, __m512i __X,
+ __m512i __Y)
+{
+ return (__m512i) __builtin_ia32_vpmadd52luq512_mask ((__v8di) __W,
+ (__v8di) __X,
+ (__v8di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask_madd52hi_epu64 (__m512i __W, __mmask8 __M, __m512i __X,
+ __m512i __Y)
+{
+ return (__m512i) __builtin_ia32_vpmadd52huq512_mask ((__v8di) __W,
+ (__v8di) __X,
+ (__v8di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_maskz_madd52lo_epu64 (__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z)
+{
+ return (__m512i) __builtin_ia32_vpmadd52luq512_maskz ((__v8di) __X,
+ (__v8di) __Y,
+ (__v8di) __Z,
+ (__mmask8) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_maskz_madd52hi_epu64 (__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z)
+{
+ return (__m512i) __builtin_ia32_vpmadd52huq512_maskz ((__v8di) __X,
+ (__v8di) __Y,
+ (__v8di) __Z,
+ (__mmask8) __M);
+}
+
+#ifdef __DISABLE_AVX512IFMA__
+#undef __DISABLE_AVX512IFMA__
+#pragma GCC pop_options
+#endif /* __DISABLE_AVX512IFMA__ */
+
+#endif /* _AVX512IFMAINTRIN_H_INCLUDED */
--- /dev/null
+/* Copyright (C) 2013-2014 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 3, 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.
+
+ Under Section 7 of GPL version 3, you are granted additional
+ permissions described in the GCC Runtime Library Exception, version
+ 3.1, as published by the Free Software Foundation.
+
+ You should have received a copy of the GNU General Public License and
+ a copy of the GCC Runtime Library Exception along with this program;
+ see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef _IMMINTRIN_H_INCLUDED
+#error "Never use <avx512ifmavlintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef _AVX512IFMAVLINTRIN_H_INCLUDED
+#define _AVX512IFMAVLINTRIN_H_INCLUDED
+
+#if !defined(__AVX512VL__) || !defined(__AVX512IFMA__)
+#pragma GCC push_options
+#pragma GCC target("avx512ifma,avx512vl")
+#define __DISABLE_AVX512IFMAVL__
+#endif /* __AVX512IFMAVL__ */
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_madd52lo_epu64 (__m128i __X, __m128i __Y, __m128i __Z)
+{
+ return (__m128i) __builtin_ia32_vpmadd52luq128_mask ((__v2di) __X,
+ (__v2di) __Y,
+ (__v2di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_madd52hi_epu64 (__m128i __X, __m128i __Y, __m128i __Z)
+{
+ return (__m128i) __builtin_ia32_vpmadd52huq128_mask ((__v2di) __X,
+ (__v2di) __Y,
+ (__v2di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_madd52lo_epu64 (__m256i __X, __m256i __Y, __m256i __Z)
+{
+ return (__m256i) __builtin_ia32_vpmadd52luq256_mask ((__v4di) __X,
+ (__v4di) __Y,
+ (__v4di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_madd52hi_epu64 (__m256i __X, __m256i __Y, __m256i __Z)
+{
+ return (__m256i) __builtin_ia32_vpmadd52huq256_mask ((__v4di) __X,
+ (__v4di) __Y,
+ (__v4di) __Z,
+ (__mmask8) - 1);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_madd52lo_epu64 (__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y)
+{
+ return (__m128i) __builtin_ia32_vpmadd52luq128_mask ((__v2di) __W,
+ (__v2di) __X,
+ (__v2di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_madd52hi_epu64 (__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y)
+{
+ return (__m128i) __builtin_ia32_vpmadd52huq128_mask ((__v2di) __W,
+ (__v2di) __X,
+ (__v2di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask_madd52lo_epu64 (__m256i __W, __mmask8 __M, __m256i __X,
+ __m256i __Y)
+{
+ return (__m256i) __builtin_ia32_vpmadd52luq256_mask ((__v4di) __W,
+ (__v4di) __X,
+ (__v4di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask_madd52hi_epu64 (__m256i __W, __mmask8 __M, __m256i __X,
+ __m256i __Y)
+{
+ return (__m256i) __builtin_ia32_vpmadd52huq256_mask ((__v4di) __W,
+ (__v4di) __X,
+ (__v4di) __Y,
+ (__mmask8) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_madd52lo_epu64 (__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z)
+{
+ return (__m128i) __builtin_ia32_vpmadd52luq128_maskz ((__v2di) __X,
+ (__v2di) __Y,
+ (__v2di) __Z,
+ (__mmask8) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_madd52hi_epu64 (__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z)
+{
+ return (__m128i) __builtin_ia32_vpmadd52huq128_maskz ((__v2di) __X,
+ (__v2di) __Y,
+ (__v2di) __Z,
+ (__mmask8) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_maskz_madd52lo_epu64 (__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z)
+{
+ return (__m256i) __builtin_ia32_vpmadd52luq256_maskz ((__v4di) __X,
+ (__v4di) __Y,
+ (__v4di) __Z,
+ (__mmask8) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_maskz_madd52hi_epu64 (__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z)
+{
+ return (__m256i) __builtin_ia32_vpmadd52huq256_maskz ((__v4di) __X,
+ (__v4di) __Y,
+ (__v4di) __Z,
+ (__mmask8) __M);
+}
+
+#ifdef __DISABLE_AVX512IFMAVL__
+#undef __DISABLE_AVX512IFMAVL__
+#pragma GCC pop_options
+#endif /* __DISABLE_AVX512IFMAVL__ */
+
+#endif /* _AVX512IFMAVLINTRIN_H_INCLUDED */
#define bit_AVX512DQ (1 << 17)
#define bit_RDSEED (1 << 18)
#define bit_ADX (1 << 19)
+#define bit_AVX512IFMA (1 << 21)
#define bit_CLFLUSHOPT (1 << 23)
#define bit_AVX512PF (1 << 26)
#define bit_AVX512ER (1 << 27)
unsigned int has_avx512f = 0, has_sha = 0, has_prefetchwt1 = 0;
unsigned int has_clflushopt = 0, has_xsavec = 0, has_xsaves = 0;
unsigned int has_avx512dq = 0, has_avx512bw = 0, has_avx512vl = 0;
+ unsigned int has_avx512ifma = 0;
bool arch;
has_avx512dq = ebx & bit_AVX512DQ;
has_avx512bw = ebx & bit_AVX512BW;
has_avx512vl = ebx & bit_AVX512VL;
+ has_avx512vl = ebx & bit_AVX512IFMA;
has_prefetchwt1 = ecx & bit_PREFETCHWT1;
}
const char *avx512dq = has_avx512dq ? " -mavx512dq" : " -mno-avx512dq";
const char *avx512bw = has_avx512bw ? " -mavx512bw" : " -mno-avx512bw";
const char *avx512vl = has_avx512vl ? " -mavx512vl" : " -mno-avx512vl";
+ const char *avx512ifma = has_avx512ifma ? " -mavx512ifma" : " -mno-avx512ifma";
options = concat (options, mmx, mmx3dnow, sse, sse2, sse3, ssse3,
sse4a, cx16, sahf, movbe, aes, sha, pclmul,
fxsr, xsave, xsaveopt, avx512f, avx512er,
avx512cd, avx512pf, prefetchwt1, clflushopt,
xsavec, xsaves, avx512dq, avx512bw, avx512vl,
- NULL);
+ avx512ifma, NULL);
}
done:
def_or_undef (parse_in, "__AVX512BW__");
if (isa_flag & OPTION_MASK_ISA_AVX512VL)
def_or_undef (parse_in, "__AVX512VL__");
+ if (isa_flag & OPTION_MASK_ISA_AVX512IFMA)
+ def_or_undef (parse_in, "__AVX512IFMA__");
if (isa_flag & OPTION_MASK_ISA_FMA)
def_or_undef (parse_in, "__FMA__");
if (isa_flag & OPTION_MASK_ISA_RTM)
{ "-mavx512dq", OPTION_MASK_ISA_AVX512DQ },
{ "-mavx512bw", OPTION_MASK_ISA_AVX512BW },
{ "-mavx512vl", OPTION_MASK_ISA_AVX512VL },
+ { "-mavx512ifma", OPTION_MASK_ISA_AVX512IFMA },
{ "-msse4a", OPTION_MASK_ISA_SSE4A },
{ "-msse4.2", OPTION_MASK_ISA_SSE4_2 },
{ "-msse4.1", OPTION_MASK_ISA_SSE4_1 },
#define PTA_AVX512DQ (HOST_WIDE_INT_1 << 50)
#define PTA_AVX512BW (HOST_WIDE_INT_1 << 51)
#define PTA_AVX512VL (HOST_WIDE_INT_1 << 52)
+#define PTA_AVX512IFMA (HOST_WIDE_INT_1 << 53)
#define PTA_CORE2 \
(PTA_64BIT | PTA_MMX | PTA_SSE | PTA_SSE2 | PTA_SSE3 | PTA_SSSE3 \
if (processor_alias_table[i].flags & PTA_MPX
&& !(opts->x_ix86_isa_flags_explicit & OPTION_MASK_ISA_MPX))
opts->x_ix86_isa_flags |= OPTION_MASK_ISA_MPX;
+ if (processor_alias_table[i].flags & PTA_AVX512IFMA
+ && !(opts->x_ix86_isa_flags_explicit & OPTION_MASK_ISA_AVX512IFMA))
+ opts->x_ix86_isa_flags |= OPTION_MASK_ISA_AVX512IFMA;
if (processor_alias_table[i].flags & (PTA_PREFETCH_SSE | PTA_SSE))
x86_prefetch_sse = true;
IX86_ATTR_ISA ("clflushopt", OPT_mclflushopt),
IX86_ATTR_ISA ("xsavec", OPT_mxsavec),
IX86_ATTR_ISA ("xsaves", OPT_mxsaves),
+ IX86_ATTR_ISA ("avx512ifma", OPT_mavx512ifma),
/* enum options */
IX86_ATTR_ENUM ("fpmath=", OPT_mfpmath_),
IX86_BUILTIN_RSQRT28SD,
IX86_BUILTIN_RSQRT28SS,
+ /* AVX-512IFMA */
+ IX86_BUILTIN_VPMADD52LUQ512,
+ IX86_BUILTIN_VPMADD52HUQ512,
+ IX86_BUILTIN_VPMADD52LUQ256,
+ IX86_BUILTIN_VPMADD52HUQ256,
+ IX86_BUILTIN_VPMADD52LUQ128,
+ IX86_BUILTIN_VPMADD52HUQ128,
+ IX86_BUILTIN_VPMADD52LUQ512_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ512_MASKZ,
+ IX86_BUILTIN_VPMADD52LUQ256_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ256_MASKZ,
+ IX86_BUILTIN_VPMADD52LUQ128_MASKZ,
+ IX86_BUILTIN_VPMADD52HUQ128_MASKZ,
+
/* SHA builtins. */
IX86_BUILTIN_SHA1MSG1,
IX86_BUILTIN_SHA1MSG2,
{ OPTION_MASK_ISA_AVX512BW, CODE_FOR_avx512bw_cmpv32hi3_mask, "__builtin_ia32_cmpw512_mask", IX86_BUILTIN_CMPW512, UNKNOWN, (int) SI_FTYPE_V32HI_V32HI_INT_SI },
{ OPTION_MASK_ISA_AVX512BW, CODE_FOR_avx512bw_ucmpv64qi3_mask, "__builtin_ia32_ucmpb512_mask", IX86_BUILTIN_UCMPB512, UNKNOWN, (int) DI_FTYPE_V64QI_V64QI_INT_DI },
{ OPTION_MASK_ISA_AVX512BW, CODE_FOR_avx512bw_ucmpv32hi3_mask, "__builtin_ia32_ucmpw512_mask", IX86_BUILTIN_UCMPW512, UNKNOWN, (int) SI_FTYPE_V32HI_V32HI_INT_SI },
+
+ /* AVX512IFMA */
+ { OPTION_MASK_ISA_AVX512IFMA, CODE_FOR_vpamdd52luqv8di_mask, "__builtin_ia32_vpmadd52luq512_mask", IX86_BUILTIN_VPMADD52LUQ512, UNKNOWN, (int) V8DI_FTYPE_V8DI_V8DI_V8DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA, CODE_FOR_vpamdd52luqv8di_maskz, "__builtin_ia32_vpmadd52luq512_maskz", IX86_BUILTIN_VPMADD52LUQ512_MASKZ, UNKNOWN, (int) V8DI_FTYPE_V8DI_V8DI_V8DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA, CODE_FOR_vpamdd52huqv8di_mask, "__builtin_ia32_vpmadd52huq512_mask", IX86_BUILTIN_VPMADD52HUQ512, UNKNOWN, (int) V8DI_FTYPE_V8DI_V8DI_V8DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA, CODE_FOR_vpamdd52huqv8di_maskz, "__builtin_ia32_vpmadd52huq512_maskz", IX86_BUILTIN_VPMADD52HUQ512_MASKZ, UNKNOWN, (int) V8DI_FTYPE_V8DI_V8DI_V8DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52luqv4di_mask, "__builtin_ia32_vpmadd52luq256_mask", IX86_BUILTIN_VPMADD52LUQ256, UNKNOWN, (int) V4DI_FTYPE_V4DI_V4DI_V4DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52luqv4di_maskz, "__builtin_ia32_vpmadd52luq256_maskz", IX86_BUILTIN_VPMADD52LUQ256_MASKZ, UNKNOWN, (int) V4DI_FTYPE_V4DI_V4DI_V4DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52huqv4di_mask, "__builtin_ia32_vpmadd52huq256_mask", IX86_BUILTIN_VPMADD52HUQ256, UNKNOWN, (int) V4DI_FTYPE_V4DI_V4DI_V4DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52huqv4di_maskz, "__builtin_ia32_vpmadd52huq256_maskz", IX86_BUILTIN_VPMADD52HUQ256_MASKZ, UNKNOWN, (int) V4DI_FTYPE_V4DI_V4DI_V4DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52luqv2di_mask, "__builtin_ia32_vpmadd52luq128_mask", IX86_BUILTIN_VPMADD52LUQ128, UNKNOWN, (int) V2DI_FTYPE_V2DI_V2DI_V2DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52luqv2di_maskz, "__builtin_ia32_vpmadd52luq128_maskz", IX86_BUILTIN_VPMADD52LUQ128_MASKZ, UNKNOWN, (int) V2DI_FTYPE_V2DI_V2DI_V2DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52huqv2di_mask, "__builtin_ia32_vpmadd52huq128_mask", IX86_BUILTIN_VPMADD52HUQ128, UNKNOWN, (int) V2DI_FTYPE_V2DI_V2DI_V2DI_QI },
+ { OPTION_MASK_ISA_AVX512IFMA | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpamdd52huqv2di_maskz, "__builtin_ia32_vpmadd52huq128_maskz", IX86_BUILTIN_VPMADD52HUQ128_MASKZ, UNKNOWN, (int) V2DI_FTYPE_V2DI_V2DI_V2DI_QI },
+
};
/* Builtins with rounding support. */
#define TARGET_AVX512BW_P(x) TARGET_ISA_AVX512BW_P(x)
#define TARGET_AVX512VL TARGET_ISA_AVX512VL
#define TARGET_AVX512VL_P(x) TARGET_ISA_AVX512VL_P(x)
+#define TARGET_AVX512IFMA TARGET_ISA_AVX512IFMA
+#define TARGET_AVX512IFMA_P(x) TARGET_ISA_AVX512IFMA_P(x)
#define TARGET_FMA TARGET_ISA_FMA
#define TARGET_FMA_P(x) TARGET_ISA_FMA_P(x)
#define TARGET_SSE4A TARGET_ISA_SSE4A
Target Report Mask(ISA_AVX512VL) Var(ix86_isa_flags) Save
Support MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AVX, AVX2 and AVX512F and AVX512VL built-in functions and code generation
+mavx512ifma
+Target Report Mask(ISA_AVX512IFMA) Var(ix86_isa_flags) Save
+Support MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AVX, AVX2 and AVX512F and AVX512IFMA built-in functions and code generation
+
mfma
Target Report Mask(ISA_FMA) Var(ix86_isa_flags) Save
Support MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AVX and FMA built-in functions and code generation
#include <avx512vldqintrin.h>
+#include <avx512ifmaintrin.h>
+
+#include <avx512ifmavlintrin.h>
+
#include <shaintrin.h>
#include <lzcntintrin.h>
UNSPEC_REDUCE
UNSPEC_FPCLASS
UNSPEC_RANGE
+
+ ;; For AVX512IFMA support
+ UNSPEC_VPMADD52LUQ
+ UNSPEC_VPMADD52HUQ
])
(define_c_enum "unspecv" [
emit_move_insn (op0, op1);
DONE;
})
+
+(define_int_iterator VPMADD52
+ [UNSPEC_VPMADD52LUQ
+ UNSPEC_VPMADD52HUQ])
+
+(define_int_attr vpmadd52type
+ [(UNSPEC_VPMADD52LUQ "luq") (UNSPEC_VPMADD52HUQ "huq")])
+
+(define_expand "vpamdd52huq<mode>_maskz"
+ [(match_operand:VI8_AVX512VL 0 "register_operand")
+ (match_operand:VI8_AVX512VL 1 "register_operand")
+ (match_operand:VI8_AVX512VL 2 "register_operand")
+ (match_operand:VI8_AVX512VL 3 "nonimmediate_operand")
+ (match_operand:<avx512fmaskmode> 4 "register_operand")]
+ "TARGET_AVX512IFMA"
+{
+ emit_insn (gen_vpamdd52huq<mode>_maskz_1 (
+ operands[0], operands[1], operands[2], operands[3],
+ CONST0_RTX (<MODE>mode), operands[4]));
+ DONE;
+})
+
+(define_expand "vpamdd52luq<mode>_maskz"
+ [(match_operand:VI8_AVX512VL 0 "register_operand")
+ (match_operand:VI8_AVX512VL 1 "register_operand")
+ (match_operand:VI8_AVX512VL 2 "register_operand")
+ (match_operand:VI8_AVX512VL 3 "nonimmediate_operand")
+ (match_operand:<avx512fmaskmode> 4 "register_operand")]
+ "TARGET_AVX512IFMA"
+{
+ emit_insn (gen_vpamdd52luq<mode>_maskz_1 (
+ operands[0], operands[1], operands[2], operands[3],
+ CONST0_RTX (<MODE>mode), operands[4]));
+ DONE;
+})
+
+(define_insn "vpamdd52<vpmadd52type><mode><sd_maskz_name>"
+ [(set (match_operand:VI8_AVX512VL 0 "register_operand" "=v")
+ (unspec:VI8_AVX512VL
+ [(match_operand:VI8_AVX512VL 1 "register_operand" "0")
+ (match_operand:VI8_AVX512VL 2 "register_operand" "v")
+ (match_operand:VI8_AVX512VL 3 "nonimmediate_operand" "vm")]
+ VPMADD52))]
+ "TARGET_AVX512IFMA"
+ "vpmadd52<vpmadd52type>\t{%3, %2, %0<sd_mask_op4>|%0<sd_mask_op4>, %2, %3}"
+ [(set_attr "type" "ssemuladd")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
+(define_insn "vpamdd52<vpmadd52type><mode>_mask"
+ [(set (match_operand:VI8_AVX512VL 0 "register_operand" "=v")
+ (vec_merge:VI8_AVX512VL
+ (unspec:VI8_AVX512VL
+ [(match_operand:VI8_AVX512VL 1 "register_operand" "0")
+ (match_operand:VI8_AVX512VL 2 "register_operand" "v")
+ (match_operand:VI8_AVX512VL 3 "nonimmediate_operand" "vm")]
+ VPMADD52)
+ (match_dup 1)
+ (match_operand:<avx512fmaskmode> 4 "register_operand" "Yk")))]
+ "TARGET_AVX512IFMA"
+ "vpmadd52<vpmadd52type>\t{%3, %2, %0%{%4%}|%0%{%4%}, %2, %3}"
+ [(set_attr "type" "ssemuladd")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
+2014-11-21 Ilya Tocar <ilya.tocar@intel.com>
+
+ * g++.dg/other/i386-2.C: Add -mavx512ifma.
+ * g++.dg/other/i386-3.C: Ditto.
+ * gcc.target/i386/avx512f-helper.h: Add avx512ifma-check.h.
+ * gcc.target/i386/avx512ifma-check.h: New.
+ * gcc.target/i386/avx512ifma-vpmaddhuq-1.c: Ditto.
+ * gcc.target/i386/avx512ifma-vpmaddhuq-2.c: Ditto.
+ * gcc.target/i386/avx512ifma-vpmaddluq-1.c: Ditto.
+ * gcc.target/i386/avx512ifma-vpmaddluq-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpmaddhuq-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpmaddluq-2.c: Ditto.
+ * gcc.target/i386/i386.exp (check_effective_target_avx512ifma): New.
+ * gcc.target/i386/sse-12.c: Add new options.
+ * gcc.target/i386/sse-13.c: Ditto.
+ * gcc.target/i386/sse-14.c: Ditto.
+ * gcc.target/i386/sse-22.c: Ditto.
+ * gcc.target/i386/sse-23.c: Ditto.
+
2014-11-21 Georg-Johann Lay <avr@gjlay.de>
Forward-port from 2014-10-30 4_9-branch r216934
/* { dg-do compile { target i?86-*-* x86_64-*-* } } */
-/* { dg-options "-O -pedantic-errors -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl" } */
+/* { dg-options "-O -pedantic-errors -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl -mavx512ifma" } */
/* Test that {,x,e,p,t,s,w,a,b,i}mmintrin.h, mm3dnow.h, fma4intrin.h,
xopintrin.h, abmintrin.h, bmiintrin.h, tbmintrin.h, lwpintrin.h,
/* { dg-do compile { target i?86-*-* x86_64-*-* } } */
-/* { dg-options "-O -fkeep-inline-functions -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl" } */
+/* { dg-options "-O -fkeep-inline-functions -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl -mavx512ifma" } */
/* Test that {,x,e,p,t,s,w,a,b,i}mmintrin.h, mm3dnow.h, fma4intrin.h,
xopintrin.h, abmintrin.h, bmiintrin.h, tbmintrin.h, lwpintrin.h,
#include "avx512bw-check.h"
#elif defined (AVX512VL)
#include "avx512vl-check.h"
+#elif defined (AVX512IFMA)
+#include "avx512ifma-check.h"
#endif
/* Macros expansion. */
#elif defined (AVX512VL)
void
avx512vl_test (void) { test_256 (); test_128 (); }
+#elif defined (AVX512IFMA)
+void
+avx512ifma_test (void) { test_512 (); }
#endif
#endif /* AVX512F_HELPER_INCLUDED */
--- /dev/null
+#include <stdlib.h>
+#include "cpuid.h"
+#include "m512-check.h"
+#include "avx512f-os-support.h"
+
+static void avx512ifma_test (void);
+
+static void __attribute__ ((noinline)) do_test (void)
+{
+ avx512ifma_test ();
+}
+
+int
+main ()
+{
+ unsigned int eax, ebx, ecx, edx;
+
+ if (!__get_cpuid (1, &eax, &ebx, &ecx, &edx))
+ return 0;
+
+ if ((ecx & bit_OSXSAVE) == (bit_OSXSAVE))
+ {
+ if (__get_cpuid_max (0, NULL) < 7)
+ return 0;
+
+ __cpuid_count (7, 0, eax, ebx, ecx, edx);
+
+ if ((avx512f_os_support ()) && ((ebx & bit_AVX512IFMA) == bit_AVX512IFMA))
+ {
+ do_test ();
+#ifdef DEBUG
+ printf ("PASSED\n");
+#endif
+ return 0;
+ }
+#ifdef DEBUG
+ printf ("SKIPPED\n");
+#endif
+ }
+#ifdef DEBUG
+ else
+ printf ("SKIPPED\n");
+#endif
+
+ return 0;
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512ifma -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52huq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+
+#include <immintrin.h>
+
+volatile __m512i _x1, _y1, _z1;
+volatile __m256i _x2, _y2, _z2;
+volatile __m128i _x3, _y3, _z3;
+
+void extern
+avx512ifma_test (void)
+{
+ _x3 = _mm_madd52hi_epu64 (_x3, _y3, _z3);
+ _x3 = _mm_mask_madd52hi_epu64 (_x3, 2, _y3, _z3);
+ _x3 = _mm_maskz_madd52hi_epu64 (2, _x3, _y3, _z3);
+ _x2 = _mm256_madd52hi_epu64 (_x2, _y2, _z2);
+ _x2 = _mm256_mask_madd52hi_epu64 (_x2, 3, _y2, _z2);
+ _x2 = _mm256_maskz_madd52hi_epu64 (3, _x2, _y2, _z2);
+ _x1 = _mm512_madd52hi_epu64 (_x1, _y1, _z1);
+ _x1 = _mm512_mask_madd52hi_epu64 (_x1, 3, _y1, _z1);
+ _x1 = _mm512_maskz_madd52hi_epu64 (3, _x1, _y1, _z1);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512ifma -DAVX512IFMA" } */
+/* { dg-require-effective-target avx512ifma } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 64)
+#include "avx512f-mask-type.h"
+
+void
+CALC (long long *r, long long *s1, long long *s2, long long *s3)
+{
+ int i;
+ long long a, b;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ /* Simulate higher 52 bits out of 104 bit,
+ by shifting opernads with 0 in lower 26 bits. */
+ a = s2[i] >> 26;
+ b = s3[i] >> 26;
+ r[i] = a * b + s1[i];
+ }
+}
+
+void
+TEST (void)
+{
+ UNION_TYPE (AVX512F_LEN, i_q) src1, src2, dst1, dst2, dst3;
+ long long dst_ref[SIZE];
+ int i;
+ MASK_TYPE mask = MASK_VALUE;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ src1.a[i] = 15 + 3467 * i;
+ src2.a[i] = 9217 + i;
+ src1.a[i] = src1.a[i] << 26;
+ src1.a[i] = src1.a[i] << 26;
+ src1.a[i] &= ((1LL << 52) - 1);
+ src2.a[i] &= ((1LL << 52) - 1);
+ dst1.a[i] = DEFAULT_VALUE;
+ dst2.a[i] = DEFAULT_VALUE;
+ dst3.a[i] = DEFAULT_VALUE;
+ }
+
+ CALC (dst_ref, dst1.a, src1.a, src2.a);
+ dst1.x = INTRINSIC (_madd52hi_epu64) (dst1.x, src1.x, src2.x);
+ dst2.x = INTRINSIC (_mask_madd52hi_epu64) (dst2.x, mask, src1.x, src2.x);
+ dst3.x = INTRINSIC (_maskz_madd52hi_epu64) (mask, dst3.x, src1.x, src2.x);
+
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst1, dst_ref))
+ abort ();
+
+ MASK_MERGE (i_q) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst2, dst_ref))
+ abort ();
+
+ MASK_ZERO (i_q) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst3, dst_ref))
+ abort ();
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512ifma -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmadd52luq\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+
+#include <immintrin.h>
+
+volatile __m512i _x1, _y1, _z1;
+volatile __m256i _x2, _y2, _z2;
+volatile __m128i _x3, _y3, _z3;
+
+void extern
+avx512ifma_test (void)
+{
+ _x3 = _mm_madd52lo_epu64 (_x3, _y3, _z3);
+ _x3 = _mm_mask_madd52lo_epu64 (_x3, 2, _y3, _z3);
+ _x3 = _mm_maskz_madd52lo_epu64 (2, _x3, _y3, _z3);
+ _x2 = _mm256_madd52lo_epu64 (_x2, _y2, _z2);
+ _x2 = _mm256_mask_madd52lo_epu64 (_x2, 3, _y2, _z2);
+ _x2 = _mm256_maskz_madd52lo_epu64 (3, _x2, _y2, _z2);
+ _x1 = _mm512_madd52lo_epu64 (_x1, _y1, _z1);
+ _x1 = _mm512_mask_madd52lo_epu64 (_x1, 3, _y1, _z1);
+ _x1 = _mm512_maskz_madd52lo_epu64 (3, _x1, _y1, _z1);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512ifma -DAVX512IFMA" } */
+/* { dg-require-effective-target avx512ifma } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 64)
+#include "avx512f-mask-type.h"
+
+void
+CALC (unsigned long long *r, unsigned long long *s1,
+ unsigned long long *s2, unsigned long long *s3)
+{
+ int i;
+
+ /* Valid, because values are less than 1 << 26. */
+ for (i = 0; i < SIZE; i++)
+ r[i] = s2[i] * s3[i] + s1[i];
+}
+
+void
+TEST (void)
+{
+ UNION_TYPE (AVX512F_LEN, i_q) src1, src2, dst1, dst2, dst3;
+ unsigned long long dst_ref[SIZE];
+ int i;
+ MASK_TYPE mask = MASK_VALUE;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ src1.a[i] = i + 50;
+ src2.a[i] = i + 100;
+ dst1.a[i] = DEFAULT_VALUE;
+ dst2.a[i] = DEFAULT_VALUE;
+ dst3.a[i] = DEFAULT_VALUE;
+ }
+
+ CALC (dst_ref, dst1.a, src1.a, src2.a);
+ dst1.x = INTRINSIC (_madd52lo_epu64) (dst1.x, src1.x, src2.x);
+ dst2.x = INTRINSIC (_mask_madd52lo_epu64) (dst2.x, mask, src1.x, src2.x);
+ dst3.x = INTRINSIC (_maskz_madd52lo_epu64) (mask, dst3.x, src1.x, src2.x);
+
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst1, dst_ref))
+ abort ();
+
+ MASK_MERGE (i_q) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst2, dst_ref))
+ abort ();
+
+ MASK_ZERO (i_q) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_q) (dst3, dst_ref))
+ abort ();
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512ifma -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512ifma-vpmaddhuq-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512ifma-vpmaddhuq-2.c"
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512ifma -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512ifma-vpmaddluq-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512ifma-vpmaddluq-2.c"
} "-mavx512bw" ]
}
+# Return 1 if avx512ifma instructions can be compiled.
+proc check_effective_target_avx512ifma { } {
+ return [check_no_compiler_messages avx512ifma object {
+ typedef long long __v8di __attribute__ ((__vector_size__ (64)));
+ __v8di
+ _mm512_madd52lo_epu64 (__v8di __X, __v8di __Y, __v8di __Z)
+ {
+ return (__v8di) __builtin_ia32_vpmadd52luq512_mask ((__v8di) __X,
+ (__v8di) __Y,
+ (__v8di) __Z,
+ -1);
+ }
+ } "-mavx512ifma" ]
+}
+
# If a testcase doesn't have special options, use these.
global DEFAULT_CFLAGS
if ![info exists DEFAULT_CFLAGS] then {
popcntintrin.h and mm_malloc.h are usable
with -O -std=c89 -pedantic-errors. */
/* { dg-do compile } */
-/* { dg-options "-O -std=c89 -pedantic-errors -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512bw -mavx512dq -mavx512vl" } */
+/* { dg-options "-O -std=c89 -pedantic-errors -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512bw -mavx512dq -mavx512vl -mavx512ifma" } */
#include <x86intrin.h>
/* { dg-do compile } */
-/* { dg-options "-O2 -Werror-implicit-function-declaration -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512vl -mavx512dq -mavx512bw" } */
+/* { dg-options "-O2 -Werror-implicit-function-declaration -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512vl -mavx512dq -mavx512bw -mavx512ifma" } */
#include <mm_malloc.h>
/* { dg-do compile } */
-/* { dg-options "-O0 -Werror-implicit-function-declaration -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl" } */
+/* { dg-options "-O0 -Werror-implicit-function-declaration -march=k8 -msse4a -m3dnow -mavx -mavx2 -mfma4 -mxop -maes -mpclmul -mpopcnt -mabm -mlzcnt -mbmi -mbmi2 -mtbm -mlwp -mfsgsbase -mrdrnd -mf16c -mfma -mrtm -mrdseed -mprfchw -madx -mfxsr -mxsaveopt -mavx512f -mavx512er -mavx512cd -mavx512pf -msha -mprefetchwt1 -mxsavec -mxsaves -mclflushopt -mavx512dq -mavx512bw -mavx512vl -mavx512ifma" } */
/* { dg-add-options bind_pic_locally } */
#include <mm_malloc.h>
#ifndef DIFFERENT_PRAGMAS
-#pragma GCC target ("sse4a,3dnow,avx,avx2,fma4,xop,aes,pclmul,popcnt,abm,lzcnt,bmi,bmi2,tbm,lwp,fsgsbase,rdrnd,f16c,rtm,rdseed,prfchw,adx,fxsr,xsaveopt,avx512f,avx512er,avx512cd,avx512pf,sha,prefetchwt1,avx512vl,avx512bw,avx512dq")
+#pragma GCC target ("sse4a,3dnow,avx,avx2,fma4,xop,aes,pclmul,popcnt,abm,lzcnt,bmi,bmi2,tbm,lwp,fsgsbase,rdrnd,f16c,rtm,rdseed,prfchw,adx,fxsr,xsaveopt,avx512f,avx512er,avx512cd,avx512pf,sha,prefetchwt1,avx512vl,avx512bw,avx512dq,avx512ifma")
#endif
/* Following intrinsics require immediate arguments. They
/* immintrin.h (AVX/AVX2/RDRND/FSGSBASE/F16C/RTM/AVX512F/SHA) */
#ifdef DIFFERENT_PRAGMAS
-#pragma GCC target ("avx,avx2,rdrnd,fsgsbase,f16c,rtm,avx512f,avx512er,avx512cd,avx512pf,sha,avx512vl,avx512bw,avx512dq")
+#pragma GCC target ("avx,avx2,rdrnd,fsgsbase,f16c,rtm,avx512f,avx512er,avx512cd,avx512pf,sha,avx512vl,avx512bw,avx512dq,avx512ifma")
#endif
#include <immintrin.h>
test_1 (_cvtss_sh, unsigned short, float, 1)
#define __builtin_ia32_extracti64x2_256_mask(A, E, C, D) __builtin_ia32_extracti64x2_256_mask(A, 1, C, D)
#define __builtin_ia32_extractf64x2_256_mask(A, E, C, D) __builtin_ia32_extractf64x2_256_mask(A, 1, C, D)
-#pragma GCC target ("sse4a,3dnow,avx,avx2,fma4,xop,aes,pclmul,popcnt,abm,lzcnt,bmi,bmi2,tbm,lwp,fsgsbase,rdrnd,f16c,fma,rtm,rdseed,prfchw,adx,fxsr,xsaveopt,avx512f,avx512er,avx512cd,avx512pf,sha,prefetchwt1,xsavec,xsaves,clflushopt,avx512bw,avx512dq,avx512vl")
+#pragma GCC target ("sse4a,3dnow,avx,avx2,fma4,xop,aes,pclmul,popcnt,abm,lzcnt,bmi,bmi2,tbm,lwp,fsgsbase,rdrnd,f16c,fma,rtm,rdseed,prfchw,adx,fxsr,xsaveopt,avx512f,avx512er,avx512cd,avx512pf,sha,prefetchwt1,xsavec,xsaves,clflushopt,avx512bw,avx512dq,avx512vl,avx512ifma")
#include <wmmintrin.h>
#include <smmintrin.h>
#include <mm3dnow.h>
projects / gcc.git / commitdiff