+2014-11-21 Ilya Tocar <ilya.tocar@intel.com>
+
+ * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VBMI_SET
+ OPTION_MASK_ISA_AVX512VBMI_UNSET): New.
+ (ix86_handle_option): Handle OPT_mavx512vbmi.
+ * config.gcc: Add avx512vbmiintrin.h, avx512vbmivlintrin.h.
+ * config/i386/avx512vbmiintrin.h: New file.
+ * config/i386/avx512vbmivlintrin.h: Ditto.
+ * config/i386/cpuid.h (bit_AVX512VBMI): New.
+ * config/i386/driver-i386.c (host_detect_local_cpu): Detect avx512vbmi.
+ * config/i386/i386-c.c (ix86_target_macros_internal): Define
+ __AVX512VBMI__.
+ * config/i386/i386.c (ix86_target_string): Add -mavx512vbmi.
+ (PTA_AVX512VBMI): Define.
+ (ix86_option_override_internal): Handle new options.
+ (ix86_valid_target_attribute_inner_p): Add avx512vbmi,
+ (ix86_builtins): Add IX86_BUILTIN_VPMULTISHIFTQB512,
+ IX86_BUILTIN_VPMULTISHIFTQB256, IX86_BUILTIN_VPMULTISHIFTQB128,
+ IX86_BUILTIN_VPERMVARQI512_MASK, IX86_BUILTIN_VPERMT2VARQI512,
+ IX86_BUILTIN_VPERMT2VARQI512_MASKZ, IX86_BUILTIN_VPERMI2VARQI512,
+ IX86_BUILTIN_VPERMVARQI256_MASK, IX86_BUILTIN_VPERMVARQI128_MASK,
+ IX86_BUILTIN_VPERMT2VARQI256, IX86_BUILTIN_VPERMT2VARQI256_MASKZ,
+ IX86_BUILTIN_VPERMT2VARQI128, IX86_BUILTIN_VPERMI2VARQI256,
+ IX86_BUILTIN_VPERMI2VARQI128.
+ (bdesc_special_args): Add __builtin_ia32_vpmultishiftqb512_mask,
+ __builtin_ia32_vpmultishiftqb256_mask,
+ __builtin_ia32_vpmultishiftqb128_mask,
+ __builtin_ia32_permvarqi512_mask, __builtin_ia32_vpermt2varqi512_mask,
+ __builtin_ia32_vpermt2varqi512_maskz,
+ __builtin_ia32_vpermi2varqi512_mask, __builtin_ia32_permvarqi256_mask,
+ __builtin_ia32_permvarqi128_mask, __builtin_ia32_vpermt2varqi256_mask,
+ __builtin_ia32_vpermt2varqi256_maskz,
+ __builtin_ia32_vpermt2varqi128_mask,
+ __builtin_ia32_vpermt2varqi128_maskz,
+ __builtin_ia32_vpermi2varqi256_mask,
+ __builtin_ia32_vpermi2varqi128_mask.
+ (ix86_hard_regno_mode_ok): Allow big masks for AVX512VBMI.
+ * config/i386/i386.h (TARGET_AVX512VBMI, TARGET_AVX512VBMI_P): Define.
+ * config/i386/i386.opt: Add mavx512vbmi.
+ * config/i386/immintrin.h: Include avx512vbmiintrin.h,
+ avx512vbmivlintrin.h.
+ * config/i386/sse.md (unspec): Add UNSPEC_VPMULTISHIFT.
+ (VI1_AVX512VL): New iterator.
+ (<avx512>_permvar<mode><mask_name>): Use it.
+ (<avx512>_vpermi2var<mode>3_maskz): Ditto.
+ (<avx512>_vpermi2var<mode>3<sd_maskz_name>): Ditto.
+ (<avx512>_vpermi2var<mode>3_mask): Ditto.
+ (<avx512>_vpermt2var<mode>3_maskz): Ditto.
+ (<avx512>_vpermt2var<mode>3<sd_maskz_name>): Ditto.
+ (<avx512>_vpermt2var<mode>3_mask): Ditto.
+ (vpmultishiftqb<mode><mask_name>): Ditto.
+
2014-11-21 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
PR rtl-optimization/63952
(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_AVX512VBMI_SET \
+ (OPTION_MASK_ISA_AVX512VBMI | 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_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_AVX512VBMI_UNSET OPTION_MASK_ISA_AVX512VBMI
#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_mavx512vbmi:
+ if (value)
+ {
+ opts->x_ix86_isa_flags |= OPTION_MASK_ISA_AVX512VBMI_SET;
+ opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_AVX512VBMI_SET;
+ }
+ else
+ {
+ opts->x_ix86_isa_flags &= ~OPTION_MASK_ISA_AVX512VBMI_UNSET;
+ opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_AVX512VBMI_UNSET;
+ }
+ return true;
+
case OPT_mfma:
if (value)
{
shaintrin.h clflushoptintrin.h xsavecintrin.h
xsavesintrin.h avx512dqintrin.h avx512bwintrin.h
avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h
- avx512ifmaintrin.h avx512ifmavlintrin.h"
+ avx512ifmaintrin.h avx512ifmavlintrin.h avx512vbmiintrin.h
+ avx512vbmivlintrin.h"
;;
x86_64-*-*)
cpu_type=i386
shaintrin.h clflushoptintrin.h xsavecintrin.h
xsavesintrin.h avx512dqintrin.h avx512bwintrin.h
avx512vlintrin.h avx512vlbwintrin.h avx512vldqintrin.h
- avx512ifmaintrin.h avx512ifmavlintrin.h"
+ avx512ifmaintrin.h avx512ifmavlintrin.h avx512vbmiintrin.h
+ avx512vbmivlintrin.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 <avx512vbmiintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef _AVX512VBMIINTRIN_H_INCLUDED
+#define _AVX512VBMIINTRIN_H_INCLUDED
+
+#ifndef __AVX512VBMI__
+#pragma GCC push_options
+#pragma GCC target("avx512vbmi")
+#define __DISABLE_AVX512VBMI__
+#endif /* __AVX512VBMI__ */
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask_multishift_epi64_epi8 (__m512i __W, __mmask64 __M, __m512i __X, __m512i __Y)
+{
+ return (__m512i) __builtin_ia32_vpmultishiftqb512_mask ((__v64qi) __X,
+ (__v64qi) __Y,
+ (__v64qi) __W,
+ (__mmask64) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_maskz_multishift_epi64_epi8 (__mmask64 __M, __m512i __X, __m512i __Y)
+{
+ return (__m512i) __builtin_ia32_vpmultishiftqb512_mask ((__v64qi) __X,
+ (__v64qi) __Y,
+ (__v64qi)
+ _mm512_setzero_si512 (),
+ (__mmask64) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_multishift_epi64_epi8 (__m512i __X, __m512i __Y)
+{
+ return (__m512i) __builtin_ia32_vpmultishiftqb512_mask ((__v64qi) __X,
+ (__v64qi) __Y,
+ (__v64qi)
+ _mm512_undefined_si512 (),
+ (__mmask64) -1);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_permutexvar_epi8 (__m512i __A, __m512i __B)
+{
+ return (__m512i) __builtin_ia32_permvarqi512_mask ((__v64qi) __B,
+ (__v64qi) __A,
+ (__v64qi)
+ _mm512_undefined_si512 (),
+ (__mmask64) -1);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_maskz_permutexvar_epi8 (__mmask64 __M, __m512i __A,
+ __m512i __B)
+{
+ return (__m512i) __builtin_ia32_permvarqi512_mask ((__v64qi) __B,
+ (__v64qi) __A,
+ (__v64qi)
+ _mm512_setzero_si512(),
+ (__mmask64) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask_permutexvar_epi8 (__m512i __W, __mmask64 __M, __m512i __A,
+ __m512i __B)
+{
+ return (__m512i) __builtin_ia32_permvarqi512_mask ((__v64qi) __B,
+ (__v64qi) __A,
+ (__v64qi) __W,
+ (__mmask64) __M);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_permutex2var_epi8 (__m512i __A, __m512i __I, __m512i __B)
+{
+ return (__m512i) __builtin_ia32_vpermt2varqi512_mask ((__v64qi) __I
+ /* idx */ ,
+ (__v64qi) __A,
+ (__v64qi) __B,
+ (__mmask64) -
+ 1);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask_permutex2var_epi8 (__m512i __A, __mmask64 __U,
+ __m512i __I, __m512i __B)
+{
+ return (__m512i) __builtin_ia32_vpermt2varqi512_mask ((__v64qi) __I
+ /* idx */ ,
+ (__v64qi) __A,
+ (__v64qi) __B,
+ (__mmask64)
+ __U);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_mask2_permutex2var_epi8 (__m512i __A, __m512i __I,
+ __mmask64 __U, __m512i __B)
+{
+ return (__m512i) __builtin_ia32_vpermi2varqi512_mask ((__v64qi) __A,
+ (__v64qi) __I
+ /* idx */ ,
+ (__v64qi) __B,
+ (__mmask64)
+ __U);
+}
+
+extern __inline __m512i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm512_maskz_permutex2var_epi8 (__mmask64 __U, __m512i __A,
+ __m512i __I, __m512i __B)
+{
+ return (__m512i) __builtin_ia32_vpermt2varqi512_maskz ((__v64qi) __I
+ /* idx */ ,
+ (__v64qi) __A,
+ (__v64qi) __B,
+ (__mmask64)
+ __U);
+}
+
+#ifdef __DISABLE_AVX512VBMI__
+#undef __DISABLE_AVX512VBMI__
+#pragma GCC pop_options
+#endif /* __DISABLE_AVX512VBMI__ */
+
+#endif /* _AVX512VBMIINTRIN_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 <avx512vbmivlintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef _AVX512VBMIVLINTRIN_H_INCLUDED
+#define _AVX512VBMIVLINTRIN_H_INCLUDED
+
+#if !defined(__AVX512VL__) || !defined(__AVX512VBMI__)
+#pragma GCC push_options
+#pragma GCC target("avx512vbmi,avx512vl")
+#define __DISABLE_AVX512VBMIVL__
+#endif /* __AVX512VBMIVL__ */
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask_multishift_epi64_epi8 (__m256i __W, __mmask32 __M, __m256i __X, __m256i __Y)
+{
+ return (__m256i) __builtin_ia32_vpmultishiftqb256_mask ((__v32qi) __X,
+ (__v32qi) __Y,
+ (__v32qi) __W,
+ (__mmask32) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_maskz_multishift_epi64_epi8 (__mmask32 __M, __m256i __X, __m256i __Y)
+{
+ return (__m256i) __builtin_ia32_vpmultishiftqb256_mask ((__v32qi) __X,
+ (__v32qi) __Y,
+ (__v32qi)
+ _mm256_setzero_si256 (),
+ (__mmask32) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_multishift_epi64_epi8 (__m256i __X, __m256i __Y)
+{
+ return (__m256i) __builtin_ia32_vpmultishiftqb256_mask ((__v32qi) __X,
+ (__v32qi) __Y,
+ (__v32qi)
+ _mm256_undefined_si256 (),
+ (__mmask32) -1);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_multishift_epi64_epi8 (__m128i __W, __mmask16 __M, __m128i __X, __m128i __Y)
+{
+ return (__m128i) __builtin_ia32_vpmultishiftqb128_mask ((__v16qi) __X,
+ (__v16qi) __Y,
+ (__v16qi) __W,
+ (__mmask16) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_multishift_epi64_epi8 (__mmask16 __M, __m128i __X, __m128i __Y)
+{
+ return (__m128i) __builtin_ia32_vpmultishiftqb128_mask ((__v16qi) __X,
+ (__v16qi) __Y,
+ (__v16qi)
+ _mm_setzero_si128 (),
+ (__mmask16) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_multishift_epi64_epi8 (__m128i __X, __m128i __Y)
+{
+ return (__m128i) __builtin_ia32_vpmultishiftqb128_mask ((__v16qi) __X,
+ (__v16qi) __Y,
+ (__v16qi)
+ _mm_undefined_si128 (),
+ (__mmask16) -1);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_permutexvar_epi8 (__m256i __A, __m256i __B)
+{
+ return (__m256i) __builtin_ia32_permvarqi256_mask ((__v32qi) __B,
+ (__v32qi) __A,
+ (__v32qi)
+ _mm256_undefined_si256 (),
+ (__mmask32) -1);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_maskz_permutexvar_epi8 (__mmask32 __M, __m256i __A,
+ __m256i __B)
+{
+ return (__m256i) __builtin_ia32_permvarqi256_mask ((__v32qi) __B,
+ (__v32qi) __A,
+ (__v32qi)
+ _mm256_setzero_si256 (),
+ (__mmask32) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask_permutexvar_epi8 (__m256i __W, __mmask32 __M, __m256i __A,
+ __m256i __B)
+{
+ return (__m256i) __builtin_ia32_permvarqi256_mask ((__v32qi) __B,
+ (__v32qi) __A,
+ (__v32qi) __W,
+ (__mmask32) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_permutexvar_epi8 (__m128i __A, __m128i __B)
+{
+ return (__m128i) __builtin_ia32_permvarqi128_mask ((__v16qi) __B,
+ (__v16qi) __A,
+ (__v16qi)
+ _mm_undefined_si128 (),
+ (__mmask16) -1);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_permutexvar_epi8 (__mmask16 __M, __m128i __A, __m128i __B)
+{
+ return (__m128i) __builtin_ia32_permvarqi128_mask ((__v16qi) __B,
+ (__v16qi) __A,
+ (__v16qi)
+ _mm_setzero_si128 (),
+ (__mmask16) __M);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_permutexvar_epi8 (__m128i __W, __mmask16 __M, __m128i __A,
+ __m128i __B)
+{
+ return (__m128i) __builtin_ia32_permvarqi128_mask ((__v16qi) __B,
+ (__v16qi) __A,
+ (__v16qi) __W,
+ (__mmask16) __M);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_permutex2var_epi8 (__m256i __A, __m256i __I, __m256i __B)
+{
+ return (__m256i) __builtin_ia32_vpermt2varqi256_mask ((__v32qi) __I
+ /* idx */ ,
+ (__v32qi) __A,
+ (__v32qi) __B,
+ (__mmask32) -
+ 1);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask_permutex2var_epi8 (__m256i __A, __mmask32 __U,
+ __m256i __I, __m256i __B)
+{
+ return (__m256i) __builtin_ia32_vpermt2varqi256_mask ((__v32qi) __I
+ /* idx */ ,
+ (__v32qi) __A,
+ (__v32qi) __B,
+ (__mmask32)
+ __U);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_mask2_permutex2var_epi8 (__m256i __A, __m256i __I,
+ __mmask32 __U, __m256i __B)
+{
+ return (__m256i) __builtin_ia32_vpermi2varqi256_mask ((__v32qi) __A,
+ (__v32qi) __I
+ /* idx */ ,
+ (__v32qi) __B,
+ (__mmask32)
+ __U);
+}
+
+extern __inline __m256i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_maskz_permutex2var_epi8 (__mmask32 __U, __m256i __A,
+ __m256i __I, __m256i __B)
+{
+ return (__m256i) __builtin_ia32_vpermt2varqi256_maskz ((__v32qi) __I
+ /* idx */ ,
+ (__v32qi) __A,
+ (__v32qi) __B,
+ (__mmask32)
+ __U);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_permutex2var_epi8 (__m128i __A, __m128i __I, __m128i __B)
+{
+ return (__m128i) __builtin_ia32_vpermt2varqi128_mask ((__v16qi) __I
+ /* idx */ ,
+ (__v16qi) __A,
+ (__v16qi) __B,
+ (__mmask16) -
+ 1);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_permutex2var_epi8 (__m128i __A, __mmask16 __U, __m128i __I,
+ __m128i __B)
+{
+ return (__m128i) __builtin_ia32_vpermt2varqi128_mask ((__v16qi) __I
+ /* idx */ ,
+ (__v16qi) __A,
+ (__v16qi) __B,
+ (__mmask16)
+ __U);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask2_permutex2var_epi8 (__m128i __A, __m128i __I, __mmask16 __U,
+ __m128i __B)
+{
+ return (__m128i) __builtin_ia32_vpermi2varqi128_mask ((__v16qi) __A,
+ (__v16qi) __I
+ /* idx */ ,
+ (__v16qi) __B,
+ (__mmask16)
+ __U);
+}
+
+extern __inline __m128i
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_permutex2var_epi8 (__mmask16 __U, __m128i __A, __m128i __I,
+ __m128i __B)
+{
+ return (__m128i) __builtin_ia32_vpermt2varqi128_maskz ((__v16qi) __I
+ /* idx */ ,
+ (__v16qi) __A,
+ (__v16qi) __B,
+ (__mmask16)
+ __U);
+}
+
+#ifdef __DISABLE_AVX512VBMIVL__
+#undef __DISABLE_AVX512VBMIVL__
+#pragma GCC pop_options
+#endif /* __DISABLE_AVX512VBMIVL__ */
+
+#endif /* _AVX512VBMIVLINTRIN_H_INCLUDED */
/* %ecx */
#define bit_PREFETCHWT1 (1 << 0)
+#define bit_AVX512VBMI (1 << 1)
/* Extended State Enumeration Sub-leaf (%eax == 13, %ecx == 1) */
#define bit_XSAVEOPT (1 << 0)
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;
+ unsigned int has_avx512vbmi = 0, has_avx512ifma = 0;
bool arch;
has_avx512vl = ebx & bit_AVX512IFMA;
has_prefetchwt1 = ecx & bit_PREFETCHWT1;
+ has_avx512vl = ecx & bit_AVX512VBMI;
}
if (max_level >= 13)
const char *avx512bw = has_avx512bw ? " -mavx512bw" : " -mno-avx512bw";
const char *avx512vl = has_avx512vl ? " -mavx512vl" : " -mno-avx512vl";
const char *avx512ifma = has_avx512ifma ? " -mavx512ifma" : " -mno-avx512ifma";
+ const char *avx512vbmi = has_avx512vbmi ? " -mavx512vbmi" : " -mno-avx512vbmi";
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,
- avx512ifma, NULL);
+ avx512ifma, avx512vbmi, 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_AVX512VBMI)
+ def_or_undef (parse_in, "__AVX512VBMI__");
if (isa_flag & OPTION_MASK_ISA_AVX512IFMA)
def_or_undef (parse_in, "__AVX512IFMA__");
if (isa_flag & OPTION_MASK_ISA_FMA)
{ "-mavx512bw", OPTION_MASK_ISA_AVX512BW },
{ "-mavx512vl", OPTION_MASK_ISA_AVX512VL },
{ "-mavx512ifma", OPTION_MASK_ISA_AVX512IFMA },
+ { "-mavx512vbmi", OPTION_MASK_ISA_AVX512VBMI },
{ "-msse4a", OPTION_MASK_ISA_SSE4A },
{ "-msse4.2", OPTION_MASK_ISA_SSE4_2 },
{ "-msse4.1", OPTION_MASK_ISA_SSE4_1 },
#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_AVX512VBMI (HOST_WIDE_INT_1 << 54)
#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_AVX512VBMI
+ && !(opts->x_ix86_isa_flags_explicit & OPTION_MASK_ISA_AVX512VBMI))
+ opts->x_ix86_isa_flags |= OPTION_MASK_ISA_AVX512VBMI;
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;
IX86_ATTR_ISA ("clflushopt", OPT_mclflushopt),
IX86_ATTR_ISA ("xsavec", OPT_mxsavec),
IX86_ATTR_ISA ("xsaves", OPT_mxsaves),
+ IX86_ATTR_ISA ("avx512vbmi", OPT_mavx512vbmi),
IX86_ATTR_ISA ("avx512ifma", OPT_mavx512ifma),
/* enum options */
IX86_BUILTIN_VPMADD52LUQ128_MASKZ,
IX86_BUILTIN_VPMADD52HUQ128_MASKZ,
+ /* AVX-512VBMI */
+ IX86_BUILTIN_VPMULTISHIFTQB512,
+ IX86_BUILTIN_VPMULTISHIFTQB256,
+ IX86_BUILTIN_VPMULTISHIFTQB128,
+ IX86_BUILTIN_VPERMVARQI512_MASK,
+ IX86_BUILTIN_VPERMT2VARQI512,
+ IX86_BUILTIN_VPERMT2VARQI512_MASKZ,
+ IX86_BUILTIN_VPERMI2VARQI512,
+ IX86_BUILTIN_VPERMVARQI256_MASK,
+ IX86_BUILTIN_VPERMVARQI128_MASK,
+ IX86_BUILTIN_VPERMT2VARQI256,
+ IX86_BUILTIN_VPERMT2VARQI256_MASKZ,
+ IX86_BUILTIN_VPERMT2VARQI128,
+ IX86_BUILTIN_VPERMT2VARQI128_MASKZ,
+ IX86_BUILTIN_VPERMI2VARQI256,
+ IX86_BUILTIN_VPERMI2VARQI128,
+
/* SHA builtins. */
IX86_BUILTIN_SHA1MSG1,
IX86_BUILTIN_SHA1MSG2,
{ 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 },
+ /* AVX512VBMI */
+ { OPTION_MASK_ISA_AVX512VBMI, CODE_FOR_vpmultishiftqbv64qi_mask, "__builtin_ia32_vpmultishiftqb512_mask", IX86_BUILTIN_VPMULTISHIFTQB512, UNKNOWN, (int) V64QI_FTYPE_V64QI_V64QI_V64QI_DI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpmultishiftqbv32qi_mask, "__builtin_ia32_vpmultishiftqb256_mask", IX86_BUILTIN_VPMULTISHIFTQB256, UNKNOWN, (int) V32QI_FTYPE_V32QI_V32QI_V32QI_SI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_vpmultishiftqbv16qi_mask, "__builtin_ia32_vpmultishiftqb128_mask", IX86_BUILTIN_VPMULTISHIFTQB128, UNKNOWN, (int) V16QI_FTYPE_V16QI_V16QI_V16QI_HI },
+ { OPTION_MASK_ISA_AVX512VBMI, CODE_FOR_avx512bw_permvarv64qi_mask, "__builtin_ia32_permvarqi512_mask", IX86_BUILTIN_VPERMVARQI512_MASK, UNKNOWN, (int) V64QI_FTYPE_V64QI_V64QI_V64QI_DI },
+ { OPTION_MASK_ISA_AVX512VBMI, CODE_FOR_avx512bw_vpermt2varv64qi3_mask, "__builtin_ia32_vpermt2varqi512_mask", IX86_BUILTIN_VPERMT2VARQI512, UNKNOWN, (int) V64QI_FTYPE_V64QI_V64QI_V64QI_DI },
+ { OPTION_MASK_ISA_AVX512VBMI, CODE_FOR_avx512bw_vpermt2varv64qi3_maskz, "__builtin_ia32_vpermt2varqi512_maskz", IX86_BUILTIN_VPERMT2VARQI512_MASKZ, UNKNOWN, (int) V64QI_FTYPE_V64QI_V64QI_V64QI_DI },
+ { OPTION_MASK_ISA_AVX512VBMI, CODE_FOR_avx512bw_vpermi2varv64qi3_mask, "__builtin_ia32_vpermi2varqi512_mask", IX86_BUILTIN_VPERMI2VARQI512, UNKNOWN, (int) V64QI_FTYPE_V64QI_V64QI_V64QI_DI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_permvarv32qi_mask, "__builtin_ia32_permvarqi256_mask", IX86_BUILTIN_VPERMVARQI256_MASK, UNKNOWN, (int) V32QI_FTYPE_V32QI_V32QI_V32QI_SI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_permvarv16qi_mask, "__builtin_ia32_permvarqi128_mask", IX86_BUILTIN_VPERMVARQI128_MASK, UNKNOWN, (int) V16QI_FTYPE_V16QI_V16QI_V16QI_HI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermt2varv32qi3_mask, "__builtin_ia32_vpermt2varqi256_mask", IX86_BUILTIN_VPERMT2VARQI256, UNKNOWN, (int) V32QI_FTYPE_V32QI_V32QI_V32QI_SI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermt2varv32qi3_maskz, "__builtin_ia32_vpermt2varqi256_maskz", IX86_BUILTIN_VPERMT2VARQI256_MASKZ, UNKNOWN, (int) V32QI_FTYPE_V32QI_V32QI_V32QI_SI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermt2varv16qi3_mask, "__builtin_ia32_vpermt2varqi128_mask", IX86_BUILTIN_VPERMT2VARQI128, UNKNOWN, (int) V16QI_FTYPE_V16QI_V16QI_V16QI_HI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermt2varv16qi3_maskz, "__builtin_ia32_vpermt2varqi128_maskz", IX86_BUILTIN_VPERMT2VARQI128_MASKZ, UNKNOWN, (int) V16QI_FTYPE_V16QI_V16QI_V16QI_HI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermi2varv32qi3_mask, "__builtin_ia32_vpermi2varqi256_mask", IX86_BUILTIN_VPERMI2VARQI256, UNKNOWN, (int) V32QI_FTYPE_V32QI_V32QI_V32QI_SI },
+ { OPTION_MASK_ISA_AVX512VBMI | OPTION_MASK_ISA_AVX512VL, CODE_FOR_avx512vl_vpermi2varv16qi3_mask, "__builtin_ia32_vpermi2varqi128_mask", IX86_BUILTIN_VPERMI2VARQI128, UNKNOWN, (int) V16QI_FTYPE_V16QI_V16QI_V16QI_HI },
};
/* Builtins with rounding support. */
return VALID_FP_MODE_P (mode);
if (MASK_REGNO_P (regno))
return (VALID_MASK_REG_MODE (mode)
- || (TARGET_AVX512BW && VALID_MASK_AVX512BW_MODE (mode)));
+ || ((TARGET_AVX512BW || TARGET_AVX512VBMI)
+ && VALID_MASK_AVX512BW_MODE (mode)));
if (BND_REGNO_P (regno))
return VALID_BND_REG_MODE (mode);
if (SSE_REGNO_P (regno))
#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_AVX512VBMI TARGET_ISA_AVX512VBMI
+#define TARGET_AVX512VBMI_P(x) TARGET_ISA_AVX512VBMI_P(x)
#define TARGET_AVX512IFMA TARGET_ISA_AVX512IFMA
#define TARGET_AVX512IFMA_P(x) TARGET_ISA_AVX512IFMA_P(x)
#define TARGET_FMA TARGET_ISA_FMA
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
+mavx512vbmi
+Target Report Mask(ISA_AVX512VBMI) Var(ix86_isa_flags) Save
+Support MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AVX, AVX2 and AVX512F and AVX512VBMI 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 <avx512ifmavlintrin.h>
+#include <avx512vbmiintrin.h>
+
+#include <avx512vbmivlintrin.h>
+
#include <shaintrin.h>
#include <lzcntintrin.h>
;; For AVX512IFMA support
UNSPEC_VPMADD52LUQ
UNSPEC_VPMADD52HUQ
+
+ ;; For AVX512VBMI support
+ UNSPEC_VPMULTISHIFT
])
(define_c_enum "unspecv" [
[V64QI (V16QI "TARGET_AVX512VL") (V32QI "TARGET_AVX512VL")
V32HI (V16HI "TARGET_AVX512VL") (V8HI "TARGET_AVX512VL")])
+(define_mode_iterator VI1_AVX512VL
+ [V64QI (V16QI "TARGET_AVX512VL") (V32QI "TARGET_AVX512VL")])
+
;; All vector modes
(define_mode_iterator V
[(V32QI "TARGET_AVX") V16QI
(set_attr "prefix" "<mask_prefix2>")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "<avx512>_permvar<mode><mask_name>"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (unspec:VI1_AVX512VL
+ [(match_operand:VI1_AVX512VL 1 "nonimmediate_operand" "vm")
+ (match_operand:<sseintvecmode> 2 "register_operand" "v")]
+ UNSPEC_VPERMVAR))]
+ "TARGET_AVX512VBMI && <mask_mode512bit_condition>"
+ "vperm<ssemodesuffix>\t{%1, %2, %0<mask_operand3>|%0<mask_operand3>, %2, %1}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "<mask_prefix2>")
+ (set_attr "mode" "<sseinsnmode>")])
+
(define_insn "<avx512>_permvar<mode><mask_name>"
[(set (match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(unspec:VI2_AVX512VL
DONE;
})
+(define_expand "<avx512>_vpermi2var<mode>3_maskz"
+ [(match_operand:VI1_AVX512VL 0 "register_operand")
+ (match_operand:VI1_AVX512VL 1 "register_operand")
+ (match_operand:<sseintvecmode> 2 "register_operand")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand")
+ (match_operand:<avx512fmaskmode> 4 "register_operand")]
+ "TARGET_AVX512VBMI"
+{
+ emit_insn (gen_<avx512>_vpermi2var<mode>3_maskz_1 (
+ operands[0], operands[1], operands[2], operands[3],
+ CONST0_RTX (<MODE>mode), operands[4]));
+ DONE;
+})
+
(define_expand "<avx512>_vpermi2var<mode>3_maskz"
[(match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(match_operand:VI2_AVX512VL 1 "register_operand" "v")
(set_attr "prefix" "evex")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "<avx512>_vpermi2var<mode>3<sd_maskz_name>"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (unspec:VI1_AVX512VL
+ [(match_operand:VI1_AVX512VL 1 "register_operand" "v")
+ (match_operand:<sseintvecmode> 2 "register_operand" "0")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand" "vm")]
+ UNSPEC_VPERMI2))]
+ "TARGET_AVX512VBMI"
+ "vpermi2<ssemodesuffix>\t{%3, %1, %0<sd_mask_op4>|%0<sd_mask_op4>, %1, %3}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
(define_insn "<avx512>_vpermi2var<mode>3<sd_maskz_name>"
[(set (match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(unspec:VI2_AVX512VL
(set_attr "prefix" "evex")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "<avx512>_vpermi2var<mode>3_mask"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (vec_merge:VI1_AVX512VL
+ (unspec:VI1_AVX512VL
+ [(match_operand:VI1_AVX512VL 1 "register_operand" "v")
+ (match_operand:<sseintvecmode> 2 "register_operand" "0")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand" "vm")]
+ UNSPEC_VPERMI2_MASK)
+ (match_dup 0)
+ (match_operand:<avx512fmaskmode> 4 "register_operand" "Yk")))]
+ "TARGET_AVX512VBMI"
+ "vpermi2<ssemodesuffix>\t{%3, %1, %0%{%4%}|%0%{%4%}, %1, %3}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
(define_insn "<avx512>_vpermi2var<mode>3_mask"
[(set (match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(vec_merge:VI2_AVX512VL
DONE;
})
+(define_expand "<avx512>_vpermt2var<mode>3_maskz"
+ [(match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (match_operand:<sseintvecmode> 1 "register_operand" "v")
+ (match_operand:VI1_AVX512VL 2 "register_operand" "0")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand" "vm")
+ (match_operand:<avx512fmaskmode> 4 "register_operand" "Yk")]
+ "TARGET_AVX512VBMI"
+{
+ emit_insn (gen_<avx512>_vpermt2var<mode>3_maskz_1 (
+ operands[0], operands[1], operands[2], operands[3],
+ CONST0_RTX (<MODE>mode), operands[4]));
+ DONE;
+})
+
(define_expand "<avx512>_vpermt2var<mode>3_maskz"
[(match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(match_operand:<sseintvecmode> 1 "register_operand" "v")
(set_attr "prefix" "evex")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "<avx512>_vpermt2var<mode>3<sd_maskz_name>"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (unspec:VI1_AVX512VL
+ [(match_operand:<sseintvecmode> 1 "register_operand" "v")
+ (match_operand:VI1_AVX512VL 2 "register_operand" "0")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand" "vm")]
+ UNSPEC_VPERMT2))]
+ "TARGET_AVX512VBMI"
+ "vpermt2<ssemodesuffix>\t{%3, %1, %0<sd_mask_op4>|%0<sd_mask_op4>, %1, %3}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
(define_insn "<avx512>_vpermt2var<mode>3<sd_maskz_name>"
[(set (match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(unspec:VI2_AVX512VL
(set_attr "prefix" "evex")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "<avx512>_vpermt2var<mode>3_mask"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (vec_merge:VI1_AVX512VL
+ (unspec:VI1_AVX512VL
+ [(match_operand:<sseintvecmode> 1 "register_operand" "v")
+ (match_operand:VI1_AVX512VL 2 "register_operand" "0")
+ (match_operand:VI1_AVX512VL 3 "nonimmediate_operand" "vm")]
+ UNSPEC_VPERMT2)
+ (match_dup 2)
+ (match_operand:<avx512fmaskmode> 4 "register_operand" "Yk")))]
+ "TARGET_AVX512VBMI"
+ "vpermt2<ssemodesuffix>\t{%3, %1, %0%{%4%}|%0%{%4%}, %1, %3}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+
(define_insn "<avx512>_vpermt2var<mode>3_mask"
[(set (match_operand:VI2_AVX512VL 0 "register_operand" "=v")
(vec_merge:VI2_AVX512VL
(set_attr "prefix" "evex")
(set_attr "mode" "<sseinsnmode>")])
+(define_insn "vpmultishiftqb<mode><mask_name>"
+ [(set (match_operand:VI1_AVX512VL 0 "register_operand" "=v")
+ (unspec:VI1_AVX512VL
+ [(match_operand:VI1_AVX512VL 1 "register_operand" "v")
+ (match_operand:VI1_AVX512VL 2 "nonimmediate_operand" "vm")]
+ UNSPEC_VPMULTISHIFT))]
+ "TARGET_AVX512VBMI"
+ "vpmultishiftqb\t{%2, %1, %0<mask_operand3>|%0<mask_operand3>, %1, %2}"
+ [(set_attr "type" "sselog")
+ (set_attr "prefix" "evex")
+ (set_attr "mode" "<sseinsnmode>")])
+2014-11-21 Ilya Tocar <ilya.tocar@intel.com>
+
+ * g++.dg/other/i386-2.C: Add -mavx512vbmi.
+ * g++.dg/other/i386-3.C: Ditto.
+ * gcc.target/i386/avx512f-helper.h: Add avx512vbmi-check.h.
+ * gcc.target/i386/avx512vbmi-check.h: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermb-1.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermb-2.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermi2b-1.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermi2b-2.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermt2b-1.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpermt2b-2.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c: Ditto.
+ * gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpermb-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpermi2b-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpermt2b-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpmaddhuq-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpmaddluq-2.c: Ditto.
+ * gcc.target/i386/avx512vl-vpmultishiftqb-2.c: Ditto.
+ * gcc.target/i386/i386.exp (check_effective_target_avx512vbmi): 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 Ilya Tocar <ilya.tocar@intel.com>
* g++.dg/other/i386-2.C: Add -mavx512ifma.
/* { 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 -mavx512ifma" } */
+/* { 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 -mavx512vbmi" } */
/* 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 -mavx512ifma" } */
+/* { 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 -mavx512vbmi" } */
/* 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 "avx512vl-check.h"
#elif defined (AVX512IFMA)
#include "avx512ifma-check.h"
+#elif defined (AVX512VBMI)
+#include "avx512vbmi-check.h"
#endif
/* Macros expansion. */
#elif defined (AVX512IFMA)
void
avx512ifma_test (void) { test_512 (); }
+#elif defined (AVX512VBMI)
+void
+avx512vbmi_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 avx512vbmi_test (void);
+
+static void __attribute__ ((noinline)) do_test (void)
+{
+ avx512vbmi_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_AVX512VBMI) == bit_AVX512VBMI))
+ {
+ 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 "-mavx512vbmi -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\{\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\{\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\{\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+
+#include <immintrin.h>
+
+volatile __m512i x1;
+volatile __m256i x2;
+volatile __m128i x3;
+volatile __mmask64 m1;
+volatile __mmask32 m2;
+volatile __mmask16 m3;
+
+void extern
+avx512bw_test (void)
+{
+ x1 = _mm512_permutexvar_epi8 (x1, x1);
+ x1 = _mm512_maskz_permutexvar_epi8 (m1, x1, x1);
+ x1 = _mm512_mask_permutexvar_epi8 (x1, m1, x1, x1);
+ x2 = _mm256_permutexvar_epi8 (x2, x2);
+ x2 = _mm256_maskz_permutexvar_epi8 (m2, x2, x2);
+ x2 = _mm256_mask_permutexvar_epi8 (x2, m2, x2, x2);
+ x3 = _mm_permutexvar_epi8 (x3, x3);
+ x3 = _mm_maskz_permutexvar_epi8 (m3, x3, x3);
+ x3 = _mm_mask_permutexvar_epi8 (x3, m3, x3, x3);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -DAVX512VBMI" } */
+/* { dg-require-effective-target avx512vbmi } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 8)
+#include "avx512f-mask-type.h"
+
+void
+CALC (char *ind, char *src, char *res)
+{
+ int i;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ res[i] = src[ind[i] & (SIZE - 1)];
+ }
+}
+
+void
+TEST (void)
+{
+ UNION_TYPE (AVX512F_LEN, i_b) s1, s2, res1, res2, res3;
+ char res_ref[SIZE];
+ MASK_TYPE mask = MASK_VALUE;
+ int i;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ s1.a[i] = i * i * i;
+ s2.a[i] = i + 20;
+ res2.a[i] = DEFAULT_VALUE;
+ }
+
+ res1.x = INTRINSIC (_permutexvar_epi8) (s1.x, s2.x);
+ res2.x = INTRINSIC (_mask_permutexvar_epi8) (res2.x, mask, s1.x, s2.x);
+ res3.x = INTRINSIC (_maskz_permutexvar_epi8) (mask, s1.x, s2.x);
+ CALC (s1.a, s2.a, res_ref);
+
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res1, res_ref))
+ abort ();
+
+ MASK_MERGE (i_b)(res_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res2, res_ref))
+ abort ();
+
+ MASK_ZERO (i_b)(res_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res3, res_ref))
+ abort ();
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512vbmi -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpermi2b\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermi2b\[ \\t\]+\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermi2b\[ \\t\]+\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+
+#include <immintrin.h>
+
+volatile __m512i x3;
+volatile __m256i x2;
+volatile __m128i x1;
+volatile __m512i z;
+volatile __m256i y;
+volatile __m128i x;
+volatile __mmask32 m3;
+volatile __mmask16 m2;
+volatile __mmask8 m1;
+
+void extern
+avx512bw_test (void)
+{
+ x3 = _mm512_mask2_permutex2var_epi8 (x3, z, m3, x3);
+ x2 = _mm256_mask2_permutex2var_epi8 (x2, y, m2, x2);
+ x1 = _mm_mask2_permutex2var_epi8 (x1, x, m1, x1);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -DAVX512VBMI" } */
+/* { dg-require-effective-target avx512vbmi } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 8)
+#include "math.h"
+#include "values.h"
+#include "avx512f-mask-type.h"
+
+#define NUM 32
+
+void
+CALC (char *dst, char *src1, char *ind, char *src2)
+{
+ int i;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ unsigned long long offset = ind[i] & (SIZE - 1);
+ unsigned long long cond = ind[i] & SIZE;
+
+ dst[i] = cond ? src2[offset] : src1[offset];
+ }
+}
+
+void
+TEST (void)
+{
+ int i, j;
+ UNION_TYPE (AVX512F_LEN, i_b) s1, s2, res, ind;
+ char res_ref[SIZE];
+
+ MASK_TYPE mask = MASK_VALUE;
+
+ for (i = 0; i < NUM; i++)
+ {
+ for (j = 0; j < SIZE; j++)
+ {
+ ind.a[j] = DEFAULT_VALUE;
+ s1.a[j] = i * 2 * j + 1;
+ s2.a[j] = i * 2 * j;
+
+ res.a[j] = DEFAULT_VALUE;
+ }
+
+ CALC (res_ref, s1.a, ind.a, s2.a);
+
+ res.x =
+ INTRINSIC (_mask2_permutex2var_epi8) (s1.x, ind.x, mask,
+ s2.x);
+
+ MASK_MERGE (i_b) (res_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res, res_ref))
+ abort ();
+ }
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512vbmi -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%zmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 3 } } *
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%xmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpermt2b\[ \\t\]+\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+
+#include <immintrin.h>
+
+volatile __m512i x3;
+volatile __m256i x2;
+volatile __m128i x1;
+volatile __m512i z;
+volatile __m256i y;
+volatile __m128i x;
+volatile __mmask32 m3;
+volatile __mmask16 m2;
+volatile __mmask8 m1;
+
+void extern
+avx512bw_test (void)
+{
+ x3 = _mm512_permutex2var_epi8 (x3, z, x3);
+ x3 = _mm512_mask_permutex2var_epi8 (x3, m3, z, x3);
+ x3 = _mm512_maskz_permutex2var_epi8 (m3, x3, z, x3);
+ x2 = _mm256_permutex2var_epi8 (x2, y, x2);
+ x2 = _mm256_mask_permutex2var_epi8 (x2, m2, y, x2);
+ x2 = _mm256_maskz_permutex2var_epi8 (m2, x2, y, x2);
+ x1 = _mm_permutex2var_epi8 (x1, x, x1);
+ x1 = _mm_mask_permutex2var_epi8 (x1, m1, x, x1);
+ x1 = _mm_maskz_permutex2var_epi8 (m1, x1, x, x1);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -DAVX512VBMI" } */
+/* { dg-require-effective-target avx512vbmi } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 8)
+#include "math.h"
+#include "values.h"
+#include "avx512f-mask-type.h"
+
+#define NUM 32
+
+void
+CALC (char *dst, char *src1, char *ind, char *src2)
+{
+ int i;
+
+ for (i = 0; i < SIZE; i++)
+ {
+ unsigned long long offset = ind[i] & (SIZE - 1);
+ unsigned long long cond = ind[i] & SIZE;
+
+ dst[i] = cond ? src2[offset] : src1[offset];
+ }
+}
+
+void
+TEST (void)
+{
+ int i, j;
+ UNION_TYPE (AVX512F_LEN, i_b) s1, s2, res1, res2, res3, ind;
+ char res_ref[SIZE];
+
+ MASK_TYPE mask = MASK_VALUE;
+
+ for (i = 0; i < NUM; i++)
+ {
+ for (j = 0; j < SIZE; j++)
+ {
+ ind.a[j] = i * (j << 1);
+ s1.a[j] = DEFAULT_VALUE;
+ s2.a[j] = 1.5 * i * 2 * j;
+
+ res1.a[j] = DEFAULT_VALUE;
+ res2.a[j] = DEFAULT_VALUE;
+ res3.a[j] = DEFAULT_VALUE;
+ }
+
+ CALC (res_ref, s1.a, ind.a, s2.a);
+
+ res1.x = INTRINSIC (_permutex2var_epi8) (s1.x, ind.x, s2.x);
+ res2.x =
+ INTRINSIC (_mask_permutex2var_epi8) (s1.x, mask, ind.x, s2.x);
+ res3.x =
+ INTRINSIC (_maskz_permutex2var_epi8) (mask, s1.x, ind.x,
+ s2.x);
+
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res1, res_ref))
+ abort ();
+
+ MASK_MERGE (i_b) (res_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res2, res_ref))
+ abort ();
+
+ MASK_ZERO (i_b) (res_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (res3, res_ref))
+ abort ();
+ }
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512vbmi -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\[^\n\]*%xmm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\[^\n\]*%ymm\[0-9\]\{%k\[1-7\]\}\{z\}" 1 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]" 3 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\[^\n\]*%zmm\[0-9\]\{%k\[1-7\]\}\[^\{\]" 1 } } */
+/* { dg-final { scan-assembler-times "vpmultishiftqb\[ \\t\]+\[^\n\]*%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
+avx512vbmi_test (void)
+{
+ _x3 = _mm_multishift_epi64_epi8 (_y3, _z3);
+ _x3 = _mm_mask_multishift_epi64_epi8 (_x3, 2, _y3, _z3);
+ _x3 = _mm_maskz_multishift_epi64_epi8 (2, _y3, _z3);
+ _x2 = _mm256_multishift_epi64_epi8 (_y2, _z2);
+ _x2 = _mm256_mask_multishift_epi64_epi8 (_x2, 3, _y2, _z2);
+ _x2 = _mm256_maskz_multishift_epi64_epi8 (3, _y2, _z2);
+ _x1 = _mm512_multishift_epi64_epi8 (_y1, _z1);
+ _x1 = _mm512_mask_multishift_epi64_epi8 (_x1, 3, _y1, _z1);
+ _x1 = _mm512_maskz_multishift_epi64_epi8 (3, _y1, _z1);
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -DAVX512VBMI" } */
+/* { dg-require-effective-target avx512vbmi } */
+
+#include "avx512f-helper.h"
+
+#define SIZE (AVX512F_LEN / 8)
+#include "avx512f-mask-type.h"
+
+void
+CALC (char *r, char *s1, char *s2)
+{
+ int i, j, k;
+ long long a, b, ctrl;
+
+ for (i = 0; i < SIZE / sizeof (long long); i++)
+ {
+ union
+ {
+ long long x;
+ char a[sizeof(long long)];
+ } src;
+
+ for (j = 0; j < sizeof (long long); j++)
+ src.a[j] = s2[i * sizeof (long long) + j];
+ for (j = 0; j < sizeof (long long); j++)
+ {
+ ctrl = s1[i * sizeof (long long) + j] & ((1 << sizeof (long long)) - 1);
+ r[i * sizeof (long long) + j] = 0;
+ for (k = 0; k < 8; k++)
+ {
+ r[i * sizeof (long long) + j] |= ((src.x >> ((ctrl + k) % (sizeof (long long) * 8))) & 1) << k;
+ }
+ }
+ }
+}
+
+void
+TEST (void)
+{
+ UNION_TYPE (AVX512F_LEN, i_b) src1, src2, dst1, dst2, dst3;
+ char 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;
+ dst2.a[i] = DEFAULT_VALUE;
+ }
+
+ CALC (dst_ref, src1.a, src2.a);
+ dst1.x = INTRINSIC (_multishift_epi64_epi8) (src1.x, src2.x);
+ dst2.x = INTRINSIC (_mask_multishift_epi64_epi8) (dst2.x, mask, src1.x, src2.x);
+ dst3.x = INTRINSIC (_maskz_multishift_epi64_epi8) (mask, src1.x, src2.x);
+
+ if (UNION_CHECK (AVX512F_LEN, i_b) (dst1, dst_ref))
+ abort ();
+
+ MASK_MERGE (i_b) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (dst2, dst_ref))
+ abort ();
+
+ MASK_ZERO (i_b) (dst_ref, mask, SIZE);
+ if (UNION_CHECK (AVX512F_LEN, i_b) (dst3, dst_ref))
+ abort ();
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermb-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermb-2.c"
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermi2b-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermi2b-2.c"
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermt2b-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpermt2b-2.c"
--- /dev/null
+/* { dg-do run } */
+/* { dg-options "-O2 -mavx512vbmi -mavx512vl -DAVX512VL" } */
+/* { dg-require-effective-target avx512vl } */
+
+#define AVX512F_LEN 256
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpmultishiftqb-2.c"
+
+#undef AVX512F_LEN
+#undef AVX512F_LEN_HALF
+
+#define AVX512F_LEN 128
+#define AVX512F_LEN_HALF 128
+#include "avx512vbmi-vpmultishiftqb-2.c"
} "-mavx512ifma" ]
}
+# Return 1 if avx512vbmi instructions can be compiled.
+proc check_effective_target_avx512vbmi { } {
+ return [check_no_compiler_messages avx512vbmi object {
+ typedef char __v64qi __attribute__ ((__vector_size__ (64)));
+ __v64qi
+ _mm512_multishift_epi64_epi8 (__v64qi __X, __v64qi __Y)
+ {
+ return (__v64qi) __builtin_ia32_vpmultishiftqb512_mask ((__v64qi) __X,
+ (__v64qi) __Y,
+ (__v64qi) __Y,
+ -1);
+ }
+ } "-mavx512vbmi" ]
+}
+
# 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 -mavx512ifma" } */
+/* { 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 -mavx512vbmi -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 -mavx512ifma" } */
+/* { 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 -mavx512vbmi -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 -mavx512ifma" } */
+/* { 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 -mavx512vbmi" } */
/* { 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,avx512ifma")
+#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,avx512vbmi,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,avx512ifma")
+#pragma GCC target ("avx,avx2,rdrnd,fsgsbase,f16c,rtm,avx512f,avx512er,avx512cd,avx512pf,sha,avx512vl,avx512bw,avx512dq,avx512ifma,avx512vbmi")
#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,avx512ifma")
+#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,avx512vbmi,avx512ifma")
#include <wmmintrin.h>
#include <smmintrin.h>
#include <mm3dnow.h>
projects / gcc.git / commitdiff