[12/22] AVX10.2 ymm rounding: Support vfmadd{132, 231, 213}p{s, d, h} intrins

Message ID	20240814090159.422097-13-haochen.jiang@intel.com
State	New
Headers	show Return-Path: <gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org> DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 4B092385B505 From: Haochen Jiang <haochen.jiang@intel.com> To: gcc-patches@gcc.gnu.org Cc: hongtao.liu@intel.com, ubizjak@gmail.com, "Hu, Lin1" <lin1.hu@intel.com> Subject: [PATCH 12/22] AVX10.2 ymm rounding: Support vfmadd{132, 231, 213}p{s, d, h} intrins Date: Wed, 14 Aug 2024 17:01:49 +0800 Message-Id: <20240814090159.422097-13-haochen.jiang@intel.com> In-Reply-To: <20240814090159.422097-1-haochen.jiang@intel.com> References: <20240814090159.422097-1-haochen.jiang@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Precedence: list Errors-To: gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org
Series	Support AVX10.2 ymm rounding \| expand [00/22] Support AVX10.2 ymm rounding [01/22] AVX10.2 ymm rounding: Support vadd{s, d, h} and vcmp{s, d, h} intrins [02/22] AVX10.2 ymm rounding: Support vcvtdq2p{s, h} and vcvtpd2p{s, h} intrins [03/22] AVX10.2 ymm rounding: Support vcvtpd2{, u}{dq, qq} intrins [04/22] AVX10.2 ymm rounding: Support vcvtph2p{s, d, sx} and vcvtph2{, u}{dq, qq} intrins [05/22] AVX10.2 ymm rounding: Support vcvtph2{, u}w and vcvtps2p{d, hx} intrins [06/22] AVX10.2 ymm rounding: Support vcvtps2{, u}{dq, qq} intrins [07/22] AVX10.2 ymm rounding: Support vcvtqq2p{s, d, h} and vcvttpd2{, u}{dq, qq} intrins [08/22] AVX10.2 ymm rounding: Support vcvttph2{, u}{dq, qq, w} intrins [09/22] AVX10.2 ymm rounding: Support vcvttps2{, u}{dq, qq} and vcvtu{dq, qq}2p{s, d, h} intrins [10/22] AVX10.2 ymm rounding: Support vcvt{, u}w2ph and vdivp{s, d, h} intrins [11/22] AVX10.2 ymm rounding: Support vfc{madd, mul}cph, vfixupimmp{s, d} intrins [12/22] AVX10.2 ymm rounding: Support vfmadd{132, 231, 213}p{s, d, h} intrins [13/22] AVX10.2 ymm rounding: Support vfmaddcph and vfmaddsub{132, 231, 213}p{s, d, h} intrins [14/22] AVX10.2 ymm rounding: Support vfm{sub, subadd}{132, 231, 213}p{s, d, h} intrins [15/22] AVX10.2 ymm rounding: Support vfmulcph and vfnmadd{132, 231, 213}p{s, d, h} intrins [16/22] AVX10.2 ymm rounding: Support vfnmsub{132, 231, 213}p{s, d, h} intrins [17/22] AVX10.2 ymm rounding: Support vgetexpp{s, d, h} and vgetmantp{s, d, h} intrins [18/22] AVX10.2 ymm rounding: Support v{max, min}p{s, d, h} intrins [19/22] AVX10.2 ymm rounding: Support vmulp{s, d, h} and vrangep{s, d} intrins [20/22] AVX10.2 ymm rounding: Support vreducep{s, d, h} and vrndscalep{s, d, h} intrins [21/22] AVX10.2 ymm rounding: Support vscalefp{s,d,h} intrins [22/22] AVX10.2 ymm rounding: Support vsqrtp{s, d, h} and vsubp{s, d, h} intrins

diff --git a/gcc/config/i386/avx10_2roundingintrin.h b/gcc/config/i386/avx10_2roundingintrin.h index d5ea6bc57da..9015095144e 100644 --- a/gcc/config/i386/avx10_2roundingintrin.h +++ b/gcc/config/i386/avx10_2roundingintrin.h @@ -2092,6 +2092,146 @@ _mm256_maskz_fixupimm_round_ps (__mmask8 __U, __m256 __A, __m256 __B, (__mmask8) __U, __R); } + +extern __inline __m256d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_fmadd_round_pd (__m256d __A, __m256d __B, __m256d __D, const int __R) +{ + return (__m256d) __builtin_ia32_vfmaddpd256_mask_round ((__v4df) __A, + (__v4df) __B, + (__v4df) __D, + (__mmask8) -1, + __R); +} + +extern __inline __m256d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask_fmadd_round_pd (__m256d __A, __mmask8 __U, __m256d __B, + __m256d __D, const int __R) +{ + return (__m256d) __builtin_ia32_vfmaddpd256_mask_round ((__v4df) __A, + (__v4df) __B, + (__v4df) __D, + (__mmask8) __U, __R); +} + +extern __inline __m256d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask3_fmadd_round_pd (__m256d __A, __m256d __B, __m256d __D, + __mmask8 __U, const int __R) +{ + return (__m256d) __builtin_ia32_vfmaddpd256_mask3_round ((__v4df) __A, + (__v4df) __B, + (__v4df) __D, + (__mmask8) __U, + __R); +} + +extern __inline __m256d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_maskz_fmadd_round_pd (__mmask8 __U, __m256d __A, __m256d __B, + __m256d __D, const int __R) +{ + return (__m256d) __builtin_ia32_vfmaddpd256_maskz_round ((__v4df) __A, + (__v4df) __B, + (__v4df) __D, + (__mmask8) __U, + __R); +} + +extern __inline __m256h +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_fmadd_round_ph (__m256h __A, __m256h __B, __m256h __D, const int __R) +{ + return (__m256h) __builtin_ia32_vfmaddph256_mask_round ((__v16hf) __A, + (__v16hf) __B, + (__v16hf) __D, + (__mmask16) -1, + __R); +} + +extern __inline __m256h +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask_fmadd_round_ph (__m256h __A, __mmask16 __U, __m256h __B, + __m256h __D, const int __R) +{ + return (__m256h) __builtin_ia32_vfmaddph256_mask_round ((__v16hf) __A, + (__v16hf) __B, + (__v16hf) __D, + (__mmask16) __U, + __R); +} + +extern __inline __m256h +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask3_fmadd_round_ph (__m256h __A, __m256h __B, __m256h __D, + __mmask16 __U, const int __R) +{ + return (__m256h) __builtin_ia32_vfmaddph256_mask3_round ((__v16hf) __A, + (__v16hf) __B, + (__v16hf) __D, + (__mmask16) __U, + __R); +} + +extern __inline __m256h +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_maskz_fmadd_round_ph (__mmask16 __U, __m256h __A, __m256h __B, + __m256h __D, const int __R) +{ + return (__m256h) __builtin_ia32_vfmaddph256_maskz_round ((__v16hf) __A, + (__v16hf) __B, + (__v16hf) __D, + (__mmask16) __U, + __R); +} + +extern __inline __m256 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_fmadd_round_ps (__m256 __A, __m256 __B, __m256 __D, const int __R) +{ + return (__m256) __builtin_ia32_vfmaddps256_mask_round ((__v8sf) __A, + (__v8sf) __B, + (__v8sf) __D, + (__mmask8) -1, + __R); +} + +extern __inline __m256 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask_fmadd_round_ps (__m256 __A, __mmask8 __U, __m256 __B, + __m256 __D, const int __R) +{ + return (__m256) __builtin_ia32_vfmaddps256_mask_round ((__v8sf) __A, + (__v8sf) __B, + (__v8sf) __D, + (__mmask8) __U, + __R); +} + +extern __inline __m256 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_mask3_fmadd_round_ps (__m256 __A, __m256 __B, __m256 __D, + __mmask8 __U, const int __R) +{ + return (__m256) __builtin_ia32_vfmaddps256_mask3_round ((__v8sf) __A, + (__v8sf) __B, + (__v8sf) __D, + (__mmask8) __U, + __R); +} + +extern __inline __m256 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm256_maskz_fmadd_round_ps (__mmask8 __U, __m256 __A, __m256 __B, + __m256 __D, const int __R) +{ + return (__m256) __builtin_ia32_vfmaddps256_maskz_round ((__v8sf) __A, + (__v8sf) __B, + (__v8sf) __D, + (__mmask8) __U, + __R); +} #else #define _mm256_add_round_pd(A, B, R) \ ((__m256d) __builtin_ia32_addpd256_mask_round ((__v4df) (A), \ @@ -3329,6 +3469,42 @@ _mm256_maskz_fixupimm_round_ps (__mmask8 __U, __m256 __A, __m256 __B, (C), \ (__mmask8) (U), \ (R))) + +#define _mm256_fmadd_round_pd(A, B, D, R) \ + ((__m256d) __builtin_ia32_vfmaddpd256_mask_round (A, B, D, -1, R)) + +#define _mm256_mask_fmadd_round_pd(A, U, B, D, R) \ + ((__m256d) __builtin_ia32_vfmaddpd256_mask_round (A, B, D, U, R)) + +#define _mm256_mask3_fmadd_round_pd(A, B, D, U, R) \ + ((__m256d) __builtin_ia32_vfmaddpd256_mask3_round (A, B, D, U, R)) + +#define _mm256_maskz_fmadd_round_pd(U, A, B, D, R) \ + ((__m256d) __builtin_ia32_vfmaddpd256_maskz_round (A, B, D, U, R)) + +#define _mm256_fmadd_round_ph(A, B, D, R) \ + ((__m256h) __builtin_ia32_vfmaddph256_mask_round (A, B, D, -1, R)) + +#define _mm256_mask_fmadd_round_ph(A, U, B, D, R) \ + ((__m256h) __builtin_ia32_vfmaddph256_mask_round (A, B, D, U, R)) + +#define _mm256_mask3_fmadd_round_ph(A, B, D, U, R) \ + ((__m256h) __builtin_ia32_vfmaddph256_mask3_round (A, B, D, U, R)) + +#define _mm256_maskz_fmadd_round_ph(U, A, B, D, R) \ + ((__m256h) __builtin_ia32_vfmaddph256_maskz_round (A, B, D, U, R)) + +#define _mm256_fmadd_round_ps(A, B, D, R) \ + ((__m256)__builtin_ia32_vfmaddps256_mask_round (A, B, D, -1, R)) + +#define _mm256_mask_fmadd_round_ps(A, U, B, D, R) \ + ((__m256)__builtin_ia32_vfmaddps256_mask_round (A, B, D, U, R)) + +#define _mm256_mask3_fmadd_round_ps(A, B, D, U, R) \ + ((__m256)__builtin_ia32_vfmaddps256_mask3_round (A, B, D, U, R)) + +#define _mm256_maskz_fmadd_round_ps(U, A, B, D, R) \ + ((__m256)__builtin_ia32_vfmaddps256_maskz_round (A, B, D, U, R)) #endif #define _mm256_cmul_round_pch(A, B, R) _mm256_fcmul_round_pch ((A), (B), (R)) diff --git a/gcc/config/i386/i386-builtin.def b/gcc/config/i386/i386-builtin.def index 55a644a67bb..33e7573503f 100644 --- a/gcc/config/i386/i386-builtin.def +++ b/gcc/config/i386/i386-builtin.def @@ -3385,6 +3385,15 @@ BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fixupimmv4df_mask_roun BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fixupimmv4df_maskz_round, "__builtin_ia32_fixupimmpd256_maskz_round", IX86_BUILTIN_VFIXUPIMMPD256_MASKZ_ROUND, UNKNOWN, (int) V4DF_FTYPE_V4DF_V4DF_V4DI_INT_UQI_INT) BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fixupimmv8sf_mask_round, "__builtin_ia32_fixupimmps256_mask_round", IX86_BUILTIN_VFIXUPIMMPS256_MASK_ROUND, UNKNOWN, (int) V8SF_FTYPE_V8SF_V8SF_V8SI_INT_UQI_INT) BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fixupimmv8sf_maskz_round, "__builtin_ia32_fixupimmps256_maskz_round", IX86_BUILTIN_VFIXUPIMMPS256_MASKZ_ROUND, UNKNOWN, (int) V8SF_FTYPE_V8SF_V8SF_V8SI_INT_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v4df_mask_round, "__builtin_ia32_vfmaddpd256_mask_round", IX86_BUILTIN_VFMADDPD256_MASK_ROUND, UNKNOWN, (int) V4DF_FTYPE_V4DF_V4DF_V4DF_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v4df_mask3_round, "__builtin_ia32_vfmaddpd256_mask3_round", IX86_BUILTIN_VFMADDPD256_MASK3_ROUND, UNKNOWN, (int) V4DF_FTYPE_V4DF_V4DF_V4DF_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v4df_maskz_round, "__builtin_ia32_vfmaddpd256_maskz_round", IX86_BUILTIN_VFMADDPD256_MASKZ_ROUND, UNKNOWN, (int) V4DF_FTYPE_V4DF_V4DF_V4DF_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v16hf_mask_round, "__builtin_ia32_vfmaddph256_mask_round", IX86_BUILTIN_VFMADDPH256_MASK_ROUND, UNKNOWN, (int) V16HF_FTYPE_V16HF_V16HF_V16HF_UHI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v16hf_mask3_round, "__builtin_ia32_vfmaddph256_mask3_round", IX86_BUILTIN_VFMADDPH512_MASK3_ROUND, UNKNOWN, (int) V16HF_FTYPE_V16HF_V16HF_V16HF_UHI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v16hf_maskz_round, "__builtin_ia32_vfmaddph256_maskz_round", IX86_BUILTIN_VFMADDPH256_MASKZ_ROUND, UNKNOWN, (int) V16HF_FTYPE_V16HF_V16HF_V16HF_UHI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v8sf_mask_round, "__builtin_ia32_vfmaddps256_mask_round", IX86_BUILTIN_VFMADDPS256_MASK_ROUND, UNKNOWN, (int) V8SF_FTYPE_V8SF_V8SF_V8SF_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v8sf_mask3_round, "__builtin_ia32_vfmaddps256_mask3_round", IX86_BUILTIN_VFMADDPS512_MASK3_ROUND, UNKNOWN, (int) V8SF_FTYPE_V8SF_V8SF_V8SF_UQI_INT) +BDESC (0, OPTION_MASK_ISA2_AVX10_2_256, CODE_FOR_avx512vl_fmadd_v8sf_maskz_round, "__builtin_ia32_vfmaddps256_maskz_round", IX86_BUILTIN_VFMADDPS256_MASKZ_ROUND, UNKNOWN, (int) V8SF_FTYPE_V8SF_V8SF_V8SF_UQI_INT) BDESC_END (ROUND_ARGS, MULTI_ARG) diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md index 10640c6fef0..fe7520ac58a 100644 --- a/gcc/config/i386/sse.md +++ b/gcc/config/i386/sse.md @@ -5738,7 +5738,7 @@ (match_operand:VFH_AVX512VL 3 "register_operand" "0")) (match_dup 3) (match_operand:<avx512fmaskmode> 4 "register_operand" "Yk")))] - "TARGET_AVX512F" + "TARGET_AVX512F && <round_mode_condition>" "vfmadd231<ssemodesuffix>\t{<round_op5>%2, %1, %0%{%4%}|%0%{%4%}, %1, %2<round_op5>}" [(set_attr "type" "ssemuladd") (set_attr "prefix" "evex") diff --git a/gcc/testsuite/gcc.target/i386/avx-1.c b/gcc/testsuite/gcc.target/i386/avx-1.c index 2e5c2f2d786..9e6e609108e 100644 --- a/gcc/testsuite/gcc.target/i386/avx-1.c +++ b/gcc/testsuite/gcc.target/i386/avx-1.c @@ -909,6 +909,15 @@ #define __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, I, E, F) __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_mask_round(A, B, C, I, E, F) __builtin_ia32_fixupimmps256_mask_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_maskz_round(A, B, C, I, E, F) __builtin_ia32_fixupimmps256_maskz_round(A, B, C, 1, E, 8) +#define __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, 8) #include <wmmintrin.h> #include <immintrin.h> diff --git a/gcc/testsuite/gcc.target/i386/avx10_2-rounding-3.c b/gcc/testsuite/gcc.target/i386/avx10_2-rounding-3.c index 16dcb274909..e5a1831147c 100644 --- a/gcc/testsuite/gcc.target/i386/avx10_2-rounding-3.c +++ b/gcc/testsuite/gcc.target/i386/avx10_2-rounding-3.c @@ -27,6 +27,18 @@ /* { dg-final { scan-assembler-times "vfixupimmps\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ /* { dg-final { scan-assembler-times "vfixupimmps\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ /* { dg-final { scan-assembler-times "vfixupimmps\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...pd\[ \\t\]+\[^\n\]*\{rn-sae\}\[^\{\n\]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...pd\[ \\t\]+\[^\n\]*\{rd-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd231pd\[ \\t\]+\[^\n\]*\{ru-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...pd\[ \\t\]+\[^\n\]*\{rz-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ph\[ \\t\]+\[^\n\]*\{rn-sae\}\[^\{\n\]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ph\[ \\t\]+\[^\n\]*\{rd-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd231ph\[ \\t\]+\[^\n\]*\{ru-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ph\[ \\t\]+\[^\n\]*\{rz-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\[^\n\]*\{rn-sae\}\[^\{\n\]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\[^\n\]*\{rd-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd231ps\[ \\t\]+\[^\n\]*\{ru-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\[^\n\]*\{rz-sae\}\[^\{\n\]*%ymm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ #include <immintrin.h> @@ -105,3 +117,22 @@ avx10_2_test_6 (void) x = _mm256_mask_fixupimm_round_ps (x, m8, x, xi, 3, _MM_FROUND_NO_EXC); x = _mm256_maskz_fixupimm_round_ps (m8, x, x, xi, 3, _MM_FROUND_NO_EXC); } + +void extern +avx10_2_test_7 (void) +{ + xd = _mm256_fmadd_round_pd (xd, xd, xd, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC); + xd = _mm256_mask_fmadd_round_pd (xd, m8, xd, xd, _MM_FROUND_TO_NEG_INF | _MM_FROUND_NO_EXC); + xd = _mm256_mask3_fmadd_round_pd (xd, xd, xd, m8, _MM_FROUND_TO_POS_INF | _MM_FROUND_NO_EXC); + xd = _mm256_maskz_fmadd_round_pd (m8, xd, xd, xd, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC); + + xh = _mm256_fmadd_round_ph (xh, xh, xh, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC); + xh = _mm256_mask_fmadd_round_ph (xh, m16, xh, xh, _MM_FROUND_TO_NEG_INF | _MM_FROUND_NO_EXC); + xh = _mm256_mask3_fmadd_round_ph (xh, xh, xh, m16, _MM_FROUND_TO_POS_INF | _MM_FROUND_NO_EXC); + xh = _mm256_maskz_fmadd_round_ph (m16, xh, xh, xh, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC); + + x = _mm256_fmadd_round_ps (x, x, x, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC); + x = _mm256_mask_fmadd_round_ps (x, m8, x, x, _MM_FROUND_TO_NEG_INF | _MM_FROUND_NO_EXC); + x = _mm256_mask3_fmadd_round_ps (x, x, x, m8, _MM_FROUND_TO_POS_INF | _MM_FROUND_NO_EXC); + x = _mm256_maskz_fmadd_round_ps (m8, x, x, x, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC); +} diff --git a/gcc/testsuite/gcc.target/i386/sse-13.c b/gcc/testsuite/gcc.target/i386/sse-13.c index b84791aeaae..d8e1ee6a99b 100644 --- a/gcc/testsuite/gcc.target/i386/sse-13.c +++ b/gcc/testsuite/gcc.target/i386/sse-13.c @@ -916,5 +916,14 @@ #define __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, D, E, F) __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_mask_round(A, B, C, D, E, F) __builtin_ia32_fixupimmps256_mask_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_maskz_round(A, B, C, D, E, F) __builtin_ia32_fixupimmps256_maskz_round(A, B, C, 1, E, 8) +#define __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, 8) #include <x86intrin.h> diff --git a/gcc/testsuite/gcc.target/i386/sse-14.c b/gcc/testsuite/gcc.target/i386/sse-14.c index c9fb7ec8345..0cf49a59a52 100644 --- a/gcc/testsuite/gcc.target/i386/sse-14.c +++ b/gcc/testsuite/gcc.target/i386/sse-14.c @@ -1178,6 +1178,9 @@ test_3 (_mm256_maskz_div_round_ph, __m256h, __mmask8, __m256h, __m256h, 9) test_3 (_mm256_maskz_div_round_ps, __m256, __mmask8, __m256, __m256, 9) test_3 (_mm256_fcmadd_round_pch, __m256h, __m256h, __m256h, __m256h, 8) test_3 (_mm256_maskz_fcmul_round_pch, __m256h, __mmask8, __m256h, __m256h, 8) +test_3 (_mm256_fmadd_round_pd, __m256d, __m256d, __m256d, __m256d, 9) +test_3 (_mm256_fmadd_round_ph, __m256h, __m256h, __m256h, __m256h, 9) +test_3 (_mm256_fmadd_round_ps, __m256, __m256, __m256, __m256, 9) test_3x (_mm256_mask_cmp_round_pd_mask, __mmask8, __mmask8, __m256d, __m256d, 1, 8) test_3x (_mm256_mask_cmp_round_ph_mask, __mmask16, __mmask16, __m256h, __m256h, 1, 8) test_3x (_mm256_mask_cmp_round_ps_mask, __mmask8, __mmask8, __m256, __m256, 1, 8) @@ -1193,6 +1196,15 @@ test_4 (_mm256_mask_fcmadd_round_pch, __m256h, __m256h, __mmask8, __m256h, __m25 test_4 (_mm256_mask3_fcmadd_round_pch, __m256h, __m256h, __m256h, __m256h, __mmask8, 9) test_4 (_mm256_maskz_fcmadd_round_pch, __m256h, __mmask8, __m256h, __m256h, __m256h, 9) test_4 (_mm256_mask_fcmul_round_pch, __m256h, __m256h, __mmask8, __m256h, __m256h, 8) +test_4 (_mm256_mask_fmadd_round_pd, __m256d, __m256d, __mmask8, __m256d, __m256d, 9) +test_4 (_mm256_mask3_fmadd_round_pd, __m256d, __m256d, __m256d, __m256d, __mmask8, 9) +test_4 (_mm256_maskz_fmadd_round_pd, __m256d, __mmask8, __m256d, __m256d, __m256d, 9) +test_4 (_mm256_mask_fmadd_round_ph, __m256h, __m256h, __mmask16, __m256h, __m256h, 9) +test_4 (_mm256_mask3_fmadd_round_ph, __m256h, __m256h, __m256h, __m256h, __mmask16, 9) +test_4 (_mm256_maskz_fmadd_round_ph, __m256h,__mmask16, __m256h, __m256h, __m256h, 9) +test_4 (_mm256_mask_fmadd_round_ps, __m256, __m256, __mmask8, __m256, __m256, 9) +test_4 (_mm256_mask3_fmadd_round_ps, __m256, __m256, __m256, __m256, __mmask8, 9) +test_4 (_mm256_maskz_fmadd_round_ps, __m256,__mmask8, __m256, __m256, __m256, 9) test_4x (_mm256_maskz_fixupimm_round_pd, __m256d, __mmask8, __m256d, __m256d, __m256i, 3, 8) test_4x (_mm256_maskz_fixupimm_round_ps, __m256, __mmask8, __m256, __m256, __m256i, 3, 8) test_4x (_mm256_mask_fixupimm_round_pd, __m256d, __m256d, __mmask8, __m256d, __m256i, 3, 8) diff --git a/gcc/testsuite/gcc.target/i386/sse-22.c b/gcc/testsuite/gcc.target/i386/sse-22.c index f9086ef03c3..9685dd072eb 100644 --- a/gcc/testsuite/gcc.target/i386/sse-22.c +++ b/gcc/testsuite/gcc.target/i386/sse-22.c @@ -1221,6 +1221,9 @@ test_3 (_mm256_maskz_div_round_ph, __m256h, __mmask8, __m256h, __m256h, 9) test_3 (_mm256_maskz_div_round_ps, __m256, __mmask8, __m256, __m256, 9) test_3 (_mm256_fcmadd_round_pch, __m256h, __m256h, __m256h, __m256h, 8) test_3 (_mm256_maskz_fcmul_round_pch, __m256h, __mmask8, __m256h, __m256h, 8) +test_3 (_mm256_fmadd_round_pd, __m256d, __m256d, __m256d, __m256d, 9) +test_3 (_mm256_fmadd_round_ph, __m256h, __m256h, __m256h, __m256h, 9) +test_3 (_mm256_fmadd_round_ps, __m256, __m256, __m256, __m256, 9) test_3x (_mm256_mask_cmp_round_pd_mask, __mmask8, __mmask8, __m256d, __m256d, 1, 8) test_3x (_mm256_mask_cmp_round_ph_mask, __mmask16, __mmask16, __m256h, __m256h, 1, 8) test_3x (_mm256_mask_cmp_round_ps_mask, __mmask8, __mmask8, __m256, __m256, 1, 8) @@ -1236,6 +1239,15 @@ test_4 (_mm256_mask_fcmadd_round_pch, __m256h, __m256h, __mmask8, __m256h, __m25 test_4 (_mm256_mask3_fcmadd_round_pch, __m256h, __m256h, __m256h, __m256h, __mmask8, 9) test_4 (_mm256_maskz_fcmadd_round_pch, __m256h, __mmask8, __m256h, __m256h, __m256h, 9) test_4 (_mm256_mask_fcmul_round_pch, __m256h, __m256h, __mmask8, __m256h, __m256h, 8) +test_4 (_mm256_mask_fmadd_round_pd, __m256d, __m256d, __mmask8, __m256d, __m256d, 9) +test_4 (_mm256_mask3_fmadd_round_pd, __m256d, __m256d, __m256d, __m256d, __mmask8, 9) +test_4 (_mm256_maskz_fmadd_round_pd, __m256d,__mmask8, __m256d, __m256d, __m256d, 9) +test_4 (_mm256_mask_fmadd_round_ph, __m256h, __m256h, __mmask16, __m256h, __m256h, 9) +test_4 (_mm256_mask3_fmadd_round_ph, __m256h, __m256h, __m256h, __m256h, __mmask16, 9) +test_4 (_mm256_maskz_fmadd_round_ph, __m256h,__mmask16, __m256h, __m256h, __m256h, 9) +test_4 (_mm256_mask_fmadd_round_ps, __m256, __m256, __mmask8, __m256, __m256, 9) +test_4 (_mm256_mask3_fmadd_round_ps, __m256, __m256, __m256, __m256, __mmask8, 9) +test_4 (_mm256_maskz_fmadd_round_ps, __m256,__mmask8, __m256, __m256, __m256, 9) test_4x (_mm256_maskz_fixupimm_round_pd, __m256d, __mmask8, __m256d, __m256d, __m256i, 3, 8) test_4x (_mm256_maskz_fixupimm_round_ps, __m256, __mmask8, __m256, __m256, __m256i, 3, 8) test_4x (_mm256_mask_fixupimm_round_pd, __m256d, __m256d, __mmask8, __m256d, __m256i, 3, 8) diff --git a/gcc/testsuite/gcc.target/i386/sse-23.c b/gcc/testsuite/gcc.target/i386/sse-23.c index 9c42d969b62..4f54afabeb5 100644 --- a/gcc/testsuite/gcc.target/i386/sse-23.c +++ b/gcc/testsuite/gcc.target/i386/sse-23.c @@ -891,6 +891,15 @@ #define __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, D, E, F) __builtin_ia32_fixupimmpd256_maskz_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_mask_round(A, B, C, D, E, F) __builtin_ia32_fixupimmps256_mask_round(A, B, C, 1, E, 8) #define __builtin_ia32_fixupimmps256_maskz_round(A, B, C, D, E, F) __builtin_ia32_fixupimmps256_maskz_round(A, B, C, 1, E, 8) +#define __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddpd256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddph256_maskz_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_mask3_round(A, B, C, D, 8) +#define __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, E) __builtin_ia32_vfmaddps256_maskz_round(A, B, C, D, 8) #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,sha,xsavec,xsaves,clflushopt,clwb,mwaitx,clzero,pku,sgx,rdpid,gfni,vpclmulqdq,pconfig,wbnoinvd,enqcmd,avx512vp2intersect,serialize,tsxldtrk,amx-tile,amx-int8,amx-bf16,kl,widekl,avxvnni,avxifma,avxvnniint8,avxneconvert,cmpccxadd,amx-fp16,prefetchi,raoint,amx-complex,avxvnniint16,sm3,sha512,sm4,avx10.2-512")

[12/22] AVX10.2 ymm rounding: Support vfmadd{132, 231, 213}p{s, d, h} intrins

Commit Message

Patch