On Fri, Nov 21, 2014 at 12:38 PM, Ilya Tocar <[email protected]> wrote:
> On 20 Nov 09:43, Uros Bizjak wrote:
>> On Wed, Nov 19, 2014 at 6:32 PM, Ilya Tocar <[email protected]> wrote:
>> > Hi,
>> >
>> > New revision of Intel ISA reference [1] has new instructions:
>> > Clwb, pcommit and new flavors of AVX512. Patch bellow adds them.
>> > I understand that stage 1 is closed, however those changes shouldn't
>> > affect anything outside if i386 backend. And are extremely unlikely to
>> > break existing functionality, and I personally think it's desirable for
>> > newest GCC to support newest spec.
>> > Bootstrapped/regtestsed on x86_64-unknown-linux-gnu.
>> > Ok for trunk?
>>
>> Please split the patch into patch series, like it was done previously
>> for AVX512F patches.
>>
>> Uros.
>>
> This part adds avx512vbmi.
> I'll send vpermi2b autogen patch together with v64qi const perm later.
> Boostraps/passes make check.
> Ok for trunk?
>
>
> gcc/
> * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VBMI_SET<
Please remove "<" in the above line.
> 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.
>
> gcc/testsuite/
>
> * 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.
OK.
Thanks,
Uros.
>
> ---
> gcc/common/config/i386/i386-common.c | 16 ++
> gcc/config.gcc | 6 +-
> gcc/config/i386/avx512vbmiintrin.h | 159 ++++++++++++
> gcc/config/i386/avx512vbmivlintrin.h | 275
> +++++++++++++++++++++
> gcc/config/i386/cpuid.h | 1 +
> gcc/config/i386/driver-i386.c | 6 +-
> gcc/config/i386/i386-c.c | 2 +
> gcc/config/i386/i386.c | 42 +++-
> gcc/config/i386/i386.h | 2 +
> gcc/config/i386/i386.opt | 4 +
> gcc/config/i386/immintrin.h | 4 +
> gcc/config/i386/sse.md | 115 +++++++++
> gcc/testsuite/g++.dg/other/i386-2.C | 2 +-
> gcc/testsuite/g++.dg/other/i386-3.C | 2 +-
> gcc/testsuite/gcc.target/i386/avx512f-helper.h | 5 +
> gcc/testsuite/gcc.target/i386/avx512vbmi-check.h | 46 ++++
> .../gcc.target/i386/avx512vbmi-vpermb-1.c | 34 +++
> .../gcc.target/i386/avx512vbmi-vpermb-2.c | 51 ++++
> .../gcc.target/i386/avx512vbmi-vpermi2b-1.c | 25 ++
> .../gcc.target/i386/avx512vbmi-vpermi2b-2.c | 58 +++++
> .../gcc.target/i386/avx512vbmi-vpermt2b-1.c | 37 +++
> .../gcc.target/i386/avx512vbmi-vpermt2b-2.c | 70 ++++++
> .../gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c | 31 +++
> .../gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c | 68 +++++
> gcc/testsuite/gcc.target/i386/avx512vl-vpermb-2.c | 14 ++
> .../gcc.target/i386/avx512vl-vpermi2b-2.c | 14 ++
> .../gcc.target/i386/avx512vl-vpermt2b-2.c | 14 ++
> .../gcc.target/i386/avx512vl-vpmultishiftqb-2.c | 14 ++
> gcc/testsuite/gcc.target/i386/i386.exp | 15 ++
> gcc/testsuite/gcc.target/i386/sse-12.c | 2 +-
> gcc/testsuite/gcc.target/i386/sse-13.c | 2 +-
> gcc/testsuite/gcc.target/i386/sse-14.c | 2 +-
> gcc/testsuite/gcc.target/i386/sse-22.c | 4 +-
> gcc/testsuite/gcc.target/i386/sse-23.c | 2 +-
> 34 files changed, 1131 insertions(+), 13 deletions(-)
> create mode 100644 gcc/config/i386/avx512vbmiintrin.h
> create mode 100644 gcc/config/i386/avx512vbmivlintrin.h
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-check.h
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-1.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-1.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-1.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-2.c
> create mode 100644
> gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c
> create mode 100644
> gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vl-vpermb-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vl-vpermi2b-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vl-vpermt2b-2.c
> create mode 100644 gcc/testsuite/gcc.target/i386/avx512vl-vpmultishiftqb-2.c
>
> diff --git a/gcc/common/config/i386/i386-common.c
> b/gcc/common/config/i386/i386-common.c
> index 73044a0..1c4f15e 100644
> --- a/gcc/common/config/i386/i386-common.c
> +++ b/gcc/common/config/i386/i386-common.c
> @@ -73,6 +73,8 @@ along with GCC; see the file COPYING3. If not see
> (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
> @@ -170,6 +172,7 @@ along with GCC; see the file COPYING3. If not see
> #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
> @@ -459,6 +462,19 @@ ix86_handle_option (struct gcc_options *opts,
> }
> 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)
> {
> diff --git a/gcc/config.gcc b/gcc/config.gcc
> index dbf4191..da2a723 100644
> --- a/gcc/config.gcc
> +++ b/gcc/config.gcc
> @@ -368,7 +368,8 @@ i[34567]86-*-*)
> 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
> @@ -388,7 +389,8 @@ x86_64-*-*)
> 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
> diff --git a/gcc/config/i386/avx512vbmiintrin.h
> b/gcc/config/i386/avx512vbmiintrin.h
> new file mode 100644
> index 0000000..c2c59ce
> --- /dev/null
> +++ b/gcc/config/i386/avx512vbmiintrin.h
> @@ -0,0 +1,159 @@
> +/* 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 */
> diff --git a/gcc/config/i386/avx512vbmivlintrin.h
> b/gcc/config/i386/avx512vbmivlintrin.h
> new file mode 100644
> index 0000000..b4ecdeb
> --- /dev/null
> +++ b/gcc/config/i386/avx512vbmivlintrin.h
> @@ -0,0 +1,275 @@
> +/* 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 */
> diff --git a/gcc/config/i386/cpuid.h b/gcc/config/i386/cpuid.h
> index e3e1ed6..0efb1a4 100644
> --- a/gcc/config/i386/cpuid.h
> +++ b/gcc/config/i386/cpuid.h
> @@ -87,6 +87,7 @@
>
> /* %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)
> diff --git a/gcc/config/i386/driver-i386.c b/gcc/config/i386/driver-i386.c
> index cb82945..72dfd04 100644
> --- a/gcc/config/i386/driver-i386.c
> +++ b/gcc/config/i386/driver-i386.c
> @@ -412,7 +412,7 @@ const char *host_detect_local_cpu (int argc, const char
> **argv)
> 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;
>
> @@ -497,6 +497,7 @@ const char *host_detect_local_cpu (int argc, const char
> **argv)
> has_avx512vl = ebx & bit_AVX512IFMA;
>
> has_prefetchwt1 = ecx & bit_PREFETCHWT1;
> + has_avx512vl = ecx & bit_AVX512VBMI;
> }
>
> if (max_level >= 13)
> @@ -928,6 +929,7 @@ const char *host_detect_local_cpu (int argc, const char
> **argv)
> 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,
> @@ -937,7 +939,7 @@ const char *host_detect_local_cpu (int argc, const char
> **argv)
> fxsr, xsave, xsaveopt, avx512f, avx512er,
> avx512cd, avx512pf, prefetchwt1, clflushopt,
> xsavec, xsaves, avx512dq, avx512bw, avx512vl,
> - avx512ifma, NULL);
> + avx512ifma, avx512vbmi, NULL);
> }
>
> done:
> diff --git a/gcc/config/i386/i386-c.c b/gcc/config/i386/i386-c.c
> index bf993d1..798eaa6 100644
> --- a/gcc/config/i386/i386-c.c
> +++ b/gcc/config/i386/i386-c.c
> @@ -351,6 +351,8 @@ ix86_target_macros_internal (HOST_WIDE_INT isa_flag,
> 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)
> diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
> index bf46258..baf3166 100644
> --- a/gcc/config/i386/i386.c
> +++ b/gcc/config/i386/i386.c
> @@ -2619,6 +2619,7 @@ ix86_target_string (HOST_WIDE_INT isa, int flags, const
> char *arch,
> { "-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 },
> @@ -3155,6 +3156,7 @@ ix86_option_override_internal (bool main_args_p,
> #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 \
> @@ -3735,6 +3737,9 @@ ix86_option_override_internal (bool main_args_p,
> 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;
> @@ -4654,6 +4659,7 @@ ix86_valid_target_attribute_inner_p (tree args, char
> *p_strings[],
> 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 */
> @@ -30064,6 +30070,23 @@ enum ix86_builtins
> 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,
> @@ -32749,6 +32772,22 @@ static const struct builtin_description bdesc_args[]
> =
> { 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. */
> @@ -41521,7 +41560,8 @@ ix86_hard_regno_mode_ok (int regno, machine_mode mode)
> 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))
> diff --git a/gcc/config/i386/i386.h b/gcc/config/i386/i386.h
> index 481d68c..2596f81 100644
> --- a/gcc/config/i386/i386.h
> +++ b/gcc/config/i386/i386.h
> @@ -77,6 +77,8 @@ see the files COPYING3 and COPYING.RUNTIME respectively.
> If not, see
> #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
> diff --git a/gcc/config/i386/i386.opt b/gcc/config/i386/i386.opt
> index 012ff8d..f6ba8a7 100644
> --- a/gcc/config/i386/i386.opt
> +++ b/gcc/config/i386/i386.opt
> @@ -657,6 +657,10 @@ 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
>
> +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
> diff --git a/gcc/config/i386/immintrin.h b/gcc/config/i386/immintrin.h
> index 5f11432..931ff15 100644
> --- a/gcc/config/i386/immintrin.h
> +++ b/gcc/config/i386/immintrin.h
> @@ -64,6 +64,10 @@
>
> #include <avx512ifmavlintrin.h>
>
> +#include <avx512vbmiintrin.h>
> +
> +#include <avx512vbmivlintrin.h>
> +
> #include <shaintrin.h>
>
> #include <lzcntintrin.h>
> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
> index 61cc904..ca5d720 100644
> --- a/gcc/config/i386/sse.md
> +++ b/gcc/config/i386/sse.md
> @@ -145,6 +145,9 @@
> ;; For AVX512IFMA support
> UNSPEC_VPMADD52LUQ
> UNSPEC_VPMADD52HUQ
> +
> + ;; For AVX512VBMI support
> + UNSPEC_VPMULTISHIFT
> ])
>
> (define_c_enum "unspecv" [
> @@ -179,6 +182,9 @@
> [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
> @@ -16469,6 +16475,18 @@
> (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
> [(match_operand:VI2_AVX512VL 1 "nonimmediate_operand" "vm")
> @@ -17011,6 +17029,20 @@
> })
>
> (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")
> (match_operand:<sseintvecmode> 2 "register_operand" "0")
> @@ -17038,6 +17070,19 @@
> (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
> [(match_operand:VI2_AVX512VL 1 "register_operand" "v")
> @@ -17067,6 +17112,22 @@
> (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
> (unspec:VI2_AVX512VL
> @@ -17097,6 +17158,20 @@
> })
>
> (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")
> (match_operand:VI2_AVX512VL 2 "register_operand" "0")
> @@ -17124,6 +17199,19 @@
> (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
> [(match_operand:<sseintvecmode> 1 "register_operand" "v")
> @@ -17153,6 +17241,22 @@
> (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
> (unspec:VI2_AVX512VL
> @@ -18519,3 +18623,14 @@
> (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>")])
> diff --git a/gcc/testsuite/g++.dg/other/i386-2.C
> b/gcc/testsuite/g++.dg/other/i386-2.C
> index a69a5e3..0368d35 100644
> --- a/gcc/testsuite/g++.dg/other/i386-2.C
> +++ b/gcc/testsuite/g++.dg/other/i386-2.C
> @@ -1,5 +1,5 @@
> /* { 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,
> diff --git a/gcc/testsuite/g++.dg/other/i386-3.C
> b/gcc/testsuite/g++.dg/other/i386-3.C
> index d3a5bbd..3a3d5ff 100644
> --- a/gcc/testsuite/g++.dg/other/i386-3.C
> +++ b/gcc/testsuite/g++.dg/other/i386-3.C
> @@ -1,5 +1,5 @@
> /* { 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,
> diff --git a/gcc/testsuite/gcc.target/i386/avx512f-helper.h
> b/gcc/testsuite/gcc.target/i386/avx512f-helper.h
> index d177429..e270cd2 100644
> --- a/gcc/testsuite/gcc.target/i386/avx512f-helper.h
> +++ b/gcc/testsuite/gcc.target/i386/avx512f-helper.h
> @@ -22,6 +22,8 @@
> #include "avx512vl-check.h"
> #elif defined (AVX512IFMA)
> #include "avx512ifma-check.h"
> +#elif defined (AVX512VBMI)
> +#include "avx512vbmi-check.h"
> #endif
>
> /* Macros expansion. */
> @@ -130,6 +132,9 @@ avx512vl_test (void) { test_256 (); test_128 (); }
> #elif defined (AVX512IFMA)
> void
> avx512ifma_test (void) { test_512 (); }
> +#elif defined (AVX512VBMI)
> +void
> +avx512vbmi_test (void) { test_512 (); }
> #endif
>
> #endif /* AVX512F_HELPER_INCLUDED */
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-check.h
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-check.h
> new file mode 100644
> index 0000000..591ff06
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-check.h
> @@ -0,0 +1,46 @@
> +#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;
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-1.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-1.c
> new file mode 100644
> index 0000000..59e568c
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-1.c
> @@ -0,0 +1,34 @@
> +/* { 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);
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-2.c
> new file mode 100644
> index 0000000..fa22fd9
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermb-2.c
> @@ -0,0 +1,51 @@
> +/* { 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 ();
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-1.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-1.c
> new file mode 100644
> index 0000000..f760c76
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-1.c
> @@ -0,0 +1,25 @@
> +/* { 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);
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-2.c
> new file mode 100644
> index 0000000..694b23b
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermi2b-2.c
> @@ -0,0 +1,58 @@
> +/* { 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 ();
> + }
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-1.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-1.c
> new file mode 100644
> index 0000000..2e67a54
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-1.c
> @@ -0,0 +1,37 @@
> +/* { 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);
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-2.c
> new file mode 100644
> index 0000000..c9f46596
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-2.c
> @@ -0,0 +1,70 @@
> +/* { 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 ();
> + }
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c
> new file mode 100644
> index 0000000..145591c
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-1.c
> @@ -0,0 +1,31 @@
> +/* { 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);
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c
> new file mode 100644
> index 0000000..936d938
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vbmi-vpmultishiftqb-2.c
> @@ -0,0 +1,68 @@
> +/* { 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 ();
> +}
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vl-vpermb-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vl-vpermb-2.c
> new file mode 100644
> index 0000000..377f34e
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vl-vpermb-2.c
> @@ -0,0 +1,14 @@
> +/* { 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"
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vl-vpermi2b-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vl-vpermi2b-2.c
> new file mode 100644
> index 0000000..bd5dfc5
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vl-vpermi2b-2.c
> @@ -0,0 +1,14 @@
> +/* { 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"
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2b-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2b-2.c
> new file mode 100644
> index 0000000..a83eeb7
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2b-2.c
> @@ -0,0 +1,14 @@
> +/* { 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"
> diff --git a/gcc/testsuite/gcc.target/i386/avx512vl-vpmultishiftqb-2.c
> b/gcc/testsuite/gcc.target/i386/avx512vl-vpmultishiftqb-2.c
> new file mode 100644
> index 0000000..d215e23
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx512vl-vpmultishiftqb-2.c
> @@ -0,0 +1,14 @@
> +/* { 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"
> diff --git a/gcc/testsuite/gcc.target/i386/i386.exp
> b/gcc/testsuite/gcc.target/i386/i386.exp
> index 060eed3..ca5ef06 100644
> --- a/gcc/testsuite/gcc.target/i386/i386.exp
> +++ b/gcc/testsuite/gcc.target/i386/i386.exp
> @@ -365,6 +365,21 @@ proc check_effective_target_avx512ifma { } {
> } "-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 {
> diff --git a/gcc/testsuite/gcc.target/i386/sse-12.c
> b/gcc/testsuite/gcc.target/i386/sse-12.c
> index 1d8fa82..a83db92 100644
> --- a/gcc/testsuite/gcc.target/i386/sse-12.c
> +++ b/gcc/testsuite/gcc.target/i386/sse-12.c
> @@ -3,7 +3,7 @@
> 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>
>
> diff --git a/gcc/testsuite/gcc.target/i386/sse-13.c
> b/gcc/testsuite/gcc.target/i386/sse-13.c
> index 878c475..f1d9157 100644
> --- a/gcc/testsuite/gcc.target/i386/sse-13.c
> +++ b/gcc/testsuite/gcc.target/i386/sse-13.c
> @@ -1,5 +1,5 @@
> /* { 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>
>
> diff --git a/gcc/testsuite/gcc.target/i386/sse-14.c
> b/gcc/testsuite/gcc.target/i386/sse-14.c
> index 4d3acb4..bc10109 100644
> --- a/gcc/testsuite/gcc.target/i386/sse-14.c
> +++ b/gcc/testsuite/gcc.target/i386/sse-14.c
> @@ -1,5 +1,5 @@
> /* { 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>
> diff --git a/gcc/testsuite/gcc.target/i386/sse-22.c
> b/gcc/testsuite/gcc.target/i386/sse-22.c
> index 7861cf5..d54d1db 100644
> --- a/gcc/testsuite/gcc.target/i386/sse-22.c
> +++ b/gcc/testsuite/gcc.target/i386/sse-22.c
> @@ -100,7 +100,7 @@
>
>
> #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
> @@ -215,7 +215,7 @@ test_4 (_mm_cmpestrz, int, __m128i, int, __m128i, int, 1)
>
> /* 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)
> diff --git a/gcc/testsuite/gcc.target/i386/sse-23.c
> b/gcc/testsuite/gcc.target/i386/sse-23.c
> index 85d403e..e699bd3 100644
> --- a/gcc/testsuite/gcc.target/i386/sse-23.c
> +++ b/gcc/testsuite/gcc.target/i386/sse-23.c
> @@ -594,7 +594,7 @@
> #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>
> --
> 1.8.3.1
>
