Author: Chaitanya Koparkar Date: 2025-09-04T14:17:05+01:00 New Revision: faca9ddce6f9480ef1e768e1df6f9f95349d15ea
URL: https://github.com/llvm/llvm-project/commit/faca9ddce6f9480ef1e768e1df6f9f95349d15ea DIFF: https://github.com/llvm/llvm-project/commit/faca9ddce6f9480ef1e768e1df6f9f95349d15ea.diff LOG: [Headers][X86] Group related AVX512VL FMA intrinsics together (NFC) (#156794) Follow-up of #156385. Added: Modified: clang/lib/Headers/avx512vlintrin.h clang/test/CodeGen/X86/avx512vl-builtins.c Removed: ################################################################################ diff --git a/clang/lib/Headers/avx512vlintrin.h b/clang/lib/Headers/avx512vlintrin.h index a85c6fdd22e5d..6780fcf0d69f2 100644 --- a/clang/lib/Headers/avx512vlintrin.h +++ b/clang/lib/Headers/avx512vlintrin.h @@ -925,6 +925,12 @@ _mm_mask_fmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { (__mmask8)__U, (__v2df)_mm_fmsub_pd(__A, __B, __C), (__v2df)__A); } +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask3_fmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + return (__m128d)__builtin_ia32_selectpd_128( + (__mmask8)__U, (__v2df)_mm_fmsub_pd(__A, __B, __C), (__v2df)__C); +} + static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_maskz_fmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { return (__m128d)__builtin_ia32_selectpd_128( @@ -932,6 +938,12 @@ _mm_maskz_fmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { (__v2df)_mm_setzero_pd()); } +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_fnmadd_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { + return (__m128d)__builtin_ia32_selectpd_128( + (__mmask8)__U, (__v2df)_mm_fnmadd_pd(__A, __B, __C), (__v2df)__A); +} + static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_mask3_fnmadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { return (__m128d)__builtin_ia32_selectpd_128( @@ -945,6 +957,18 @@ _mm_maskz_fnmadd_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { (__v2df)_mm_setzero_pd()); } +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_fnmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { + return (__m128d)__builtin_ia32_selectpd_128( + (__mmask8)__U, (__v2df)_mm_fnmsub_pd(__A, __B, __C), (__v2df)__A); +} + +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask3_fnmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + return (__m128d)__builtin_ia32_selectpd_128( + (__mmask8)__U, (__v2df)_mm_fnmsub_pd(__A, __B, __C), (__v2df)__C); +} + static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_maskz_fnmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { return (__m128d)__builtin_ia32_selectpd_128( @@ -977,6 +1001,12 @@ _mm256_mask_fmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { (__mmask8)__U, (__v4df)_mm256_fmsub_pd(__A, __B, __C), (__v4df)__A); } +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask3_fmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + return (__m256d)__builtin_ia32_selectpd_256( + (__mmask8)__U, (__v4df)_mm256_fmsub_pd(__A, __B, __C), (__v4df)__C); +} + static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_maskz_fmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { return (__m256d)__builtin_ia32_selectpd_256( @@ -984,6 +1014,12 @@ _mm256_maskz_fmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { (__v4df)_mm256_setzero_pd()); } +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_fnmadd_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { + return (__m256d)__builtin_ia32_selectpd_256( + (__mmask8)__U, (__v4df)_mm256_fnmadd_pd(__A, __B, __C), (__v4df)__A); +} + static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mask3_fnmadd_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { return (__m256d)__builtin_ia32_selectpd_256( @@ -997,6 +1033,18 @@ _mm256_maskz_fnmadd_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { (__v4df)_mm256_setzero_pd()); } +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_fnmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { + return (__m256d)__builtin_ia32_selectpd_256( + (__mmask8)__U, (__v4df)_mm256_fnmsub_pd(__A, __B, __C), (__v4df)__A); +} + +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask3_fnmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + return (__m256d)__builtin_ia32_selectpd_256( + (__mmask8)__U, (__v4df)_mm256_fnmsub_pd(__A, __B, __C), (__v4df)__C); +} + static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_maskz_fnmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { return (__m256d)__builtin_ia32_selectpd_256( @@ -1029,6 +1077,12 @@ _mm_mask_fmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { (__mmask8)__U, (__v4sf)_mm_fmsub_ps(__A, __B, __C), (__v4sf)__A); } +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask3_fmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + return (__m128)__builtin_ia32_selectps_128( + (__mmask8)__U, (__v4sf)_mm_fmsub_ps(__A, __B, __C), (__v4sf)__C); +} + static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_maskz_fmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { return (__m128)__builtin_ia32_selectps_128( @@ -1036,6 +1090,12 @@ _mm_maskz_fmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { (__v4sf)_mm_setzero_ps()); } +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_fnmadd_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { + return (__m128)__builtin_ia32_selectps_128( + (__mmask8)__U, (__v4sf)_mm_fnmadd_ps(__A, __B, __C), (__v4sf)__A); +} + static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_mask3_fnmadd_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { return (__m128)__builtin_ia32_selectps_128( @@ -1049,6 +1109,18 @@ _mm_maskz_fnmadd_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { (__v4sf)_mm_setzero_ps()); } +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_fnmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { + return (__m128)__builtin_ia32_selectps_128( + (__mmask8)__U, (__v4sf)_mm_fnmsub_ps(__A, __B, __C), (__v4sf)__A); +} + +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask3_fnmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + return (__m128)__builtin_ia32_selectps_128( + (__mmask8)__U, (__v4sf)_mm_fnmsub_ps(__A, __B, __C), (__v4sf)__C); +} + static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_maskz_fnmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { return (__m128)__builtin_ia32_selectps_128( @@ -1081,6 +1153,12 @@ _mm256_mask_fmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { (__mmask8)__U, (__v8sf)_mm256_fmsub_ps(__A, __B, __C), (__v8sf)__A); } +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask3_fmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + return (__m256)__builtin_ia32_selectps_256( + (__mmask8)__U, (__v8sf)_mm256_fmsub_ps(__A, __B, __C), (__v8sf)__C); +} + static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_maskz_fmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { return (__m256)__builtin_ia32_selectps_256( @@ -1088,6 +1166,12 @@ _mm256_maskz_fmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { (__v8sf)_mm256_setzero_ps()); } +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_fnmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { + return (__m256)__builtin_ia32_selectps_256( + (__mmask8)__U, (__v8sf)_mm256_fnmadd_ps(__A, __B, __C), (__v8sf)__A); +} + static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mask3_fnmadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { return (__m256)__builtin_ia32_selectps_256( @@ -1101,6 +1185,18 @@ _mm256_maskz_fnmadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { (__v8sf)_mm256_setzero_ps()); } +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_fnmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { + return (__m256)__builtin_ia32_selectps_256( + (__mmask8)__U, (__v8sf)_mm256_fnmsub_ps(__A, __B, __C), (__v8sf)__A); +} + +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask3_fnmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + return (__m256)__builtin_ia32_selectps_256( + (__mmask8)__U, (__v8sf)_mm256_fnmsub_ps(__A, __B, __C), (__v8sf)__C); +} + static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_maskz_fnmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { return (__m256)__builtin_ia32_selectps_256( @@ -1308,30 +1404,6 @@ _mm256_maskz_fmsubadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) (__v8sf)_mm256_setzero_ps()); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask3_fmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - return (__m128d)__builtin_ia32_selectpd_128( - (__mmask8)__U, (__v2df)_mm_fmsub_pd(__A, __B, __C), (__v2df)__C); -} - -static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask3_fmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - return (__m256d)__builtin_ia32_selectpd_256( - (__mmask8)__U, (__v4df)_mm256_fmsub_pd(__A, __B, __C), (__v4df)__C); -} - -static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask3_fmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - return (__m128)__builtin_ia32_selectps_128( - (__mmask8)__U, (__v4sf)_mm_fmsub_ps(__A, __B, __C), (__v4sf)__C); -} - -static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask3_fmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - return (__m256)__builtin_ia32_selectps_256( - (__mmask8)__U, (__v8sf)_mm256_fmsub_ps(__A, __B, __C), (__v8sf)__C); -} - static __inline__ __m128d __DEFAULT_FN_ATTRS128 _mm_mask3_fmsubadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { @@ -1372,78 +1444,6 @@ _mm256_mask3_fmsubadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) (__v8sf) __C); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask_fnmadd_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { - return (__m128d)__builtin_ia32_selectpd_128( - (__mmask8)__U, (__v2df)_mm_fnmadd_pd(__A, __B, __C), (__v2df)__A); -} - -static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask_fnmadd_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { - return (__m256d)__builtin_ia32_selectpd_256( - (__mmask8)__U, (__v4df)_mm256_fnmadd_pd(__A, __B, __C), (__v4df)__A); -} - -static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask_fnmadd_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { - return (__m128)__builtin_ia32_selectps_128( - (__mmask8)__U, (__v4sf)_mm_fnmadd_ps(__A, __B, __C), (__v4sf)__A); -} - -static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask_fnmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { - return (__m256)__builtin_ia32_selectps_256( - (__mmask8)__U, (__v8sf)_mm256_fnmadd_ps(__A, __B, __C), (__v8sf)__A); -} - -static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask_fnmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { - return (__m128d)__builtin_ia32_selectpd_128( - (__mmask8)__U, (__v2df)_mm_fnmsub_pd(__A, __B, __C), (__v2df)__A); -} - -static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask3_fnmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - return (__m128d)__builtin_ia32_selectpd_128( - (__mmask8)__U, (__v2df)_mm_fnmsub_pd(__A, __B, __C), (__v2df)__C); -} - -static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask_fnmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { - return (__m256d)__builtin_ia32_selectpd_256( - (__mmask8)__U, (__v4df)_mm256_fnmsub_pd(__A, __B, __C), (__v4df)__A); -} - -static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask3_fnmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - return (__m256d)__builtin_ia32_selectpd_256( - (__mmask8)__U, (__v4df)_mm256_fnmsub_pd(__A, __B, __C), (__v4df)__C); -} - -static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask_fnmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { - return (__m128)__builtin_ia32_selectps_128( - (__mmask8)__U, (__v4sf)_mm_fnmsub_ps(__A, __B, __C), (__v4sf)__A); -} - -static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_mask3_fnmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - return (__m128)__builtin_ia32_selectps_128( - (__mmask8)__U, (__v4sf)_mm_fnmsub_ps(__A, __B, __C), (__v4sf)__C); -} - -static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask_fnmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { - return (__m256)__builtin_ia32_selectps_256( - (__mmask8)__U, (__v8sf)_mm256_fnmsub_ps(__A, __B, __C), (__v8sf)__A); -} - -static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR -_mm256_mask3_fnmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - return (__m256)__builtin_ia32_selectps_256( - (__mmask8)__U, (__v8sf)_mm256_fnmsub_ps(__A, __B, __C), (__v8sf)__C); -} - static __inline__ __m128d __DEFAULT_FN_ATTRS128 _mm_mask_add_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) { return (__m128d)__builtin_ia32_selectpd_128((__mmask8)__U, diff --git a/clang/test/CodeGen/X86/avx512vl-builtins.c b/clang/test/CodeGen/X86/avx512vl-builtins.c index ab5d80c973242..fb0e18662de3c 100644 --- a/clang/test/CodeGen/X86/avx512vl-builtins.c +++ b/clang/test/CodeGen/X86/avx512vl-builtins.c @@ -2853,6 +2853,26 @@ __m128d test_mm_mask_fmadd_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __ TEST_CONSTEXPR(match_m128d(_mm_mask_fmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, 14.0)); TEST_CONSTEXPR(match_m128d(_mm_mask_fmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 8.0, 2.0)); +__m128d test_mm_mask3_fmadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fmadd_pd + // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> + // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} + return _mm_mask3_fmadd_pd(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m128d(_mm_mask3_fmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, 14.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), 8.0, 6.0)); + +__m128d test_mm_maskz_fmadd_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { + // CHECK-LABEL: test_mm_maskz_fmadd_pd + // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> + // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} + return _mm_maskz_fmadd_pd(__U, __A, __B, __C); +} +TEST_CONSTEXPR(match_m128d(_mm_maskz_fmadd_pd((__mmask8)0b10, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 0.0, 14.0)); +TEST_CONSTEXPR(match_m128d(_mm_maskz_fmadd_pd((__mmask8)0b01, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 8.0, 0.0)); + __m128d test_mm_mask_fmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { // CHECK-LABEL: test_mm_mask_fmsub_pd // CHECK: fneg <2 x double> %{{.*}} @@ -2864,47 +2884,49 @@ __m128d test_mm_mask_fmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __ TEST_CONSTEXPR(match_m128d(_mm_mask_fmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, 2.0)); TEST_CONSTEXPR(match_m128d(_mm_mask_fmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), -2.0, 2.0)); -__m128d test_mm_mask3_fmadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fmadd_pd +__m128d test_mm_mask3_fmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fmsub_pd + // CHECK: fneg <2 x double> %{{.*}} // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask3_fmadd_pd(__A, __B, __C, __U); + return _mm_mask3_fmsub_pd(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m128d(_mm_mask3_fmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, 14.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask3_fmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), 8.0, 6.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, 2.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), -2.0, 6.0)); -__m128d test_mm_mask3_fnmadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fnmadd_pd +__m128d test_mm_maskz_fmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { + // CHECK-LABEL: test_mm_maskz_fmsub_pd // CHECK: fneg <2 x double> %{{.*}} // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask3_fnmadd_pd(__A, __B, __C, __U); + return _mm_maskz_fmsub_pd(__U, __A, __B, __C); } -TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, -2.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), 2.0, 6.0)); +TEST_CONSTEXPR(match_m128d(_mm_maskz_fmsub_pd((__mmask8)0b10, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 0.0, 2.0)); +TEST_CONSTEXPR(match_m128d(_mm_maskz_fmsub_pd((__mmask8)0b01, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), -2.0, 0.0)); -__m128d test_mm_maskz_fmadd_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { - // CHECK-LABEL: test_mm_maskz_fmadd_pd +__m128d test_mm_mask_fnmadd_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { + // CHECK-LABEL: test_mm_mask_fnmadd_pd + // CHECK: fneg <2 x double> %{{.*}} // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_maskz_fmadd_pd(__U, __A, __B, __C); + return _mm_mask_fnmadd_pd(__A, __U, __B, __C); } -TEST_CONSTEXPR(match_m128d(_mm_maskz_fmadd_pd((__mmask8)0b10, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 0.0, 14.0)); -TEST_CONSTEXPR(match_m128d(_mm_maskz_fmadd_pd((__mmask8)0b01, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 8.0, 0.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask_fnmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, -2.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask_fnmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 2.0, 2.0)); -__m128d test_mm_maskz_fmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { - // CHECK-LABEL: test_mm_maskz_fmsub_pd +__m128d test_mm_mask3_fnmadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fnmadd_pd // CHECK: fneg <2 x double> %{{.*}} // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_maskz_fmsub_pd(__U, __A, __B, __C); + return _mm_mask3_fnmadd_pd(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m128d(_mm_maskz_fmsub_pd((__mmask8)0b10, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 0.0, 2.0)); -TEST_CONSTEXPR(match_m128d(_mm_maskz_fmsub_pd((__mmask8)0b01, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), -2.0, 0.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, -2.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmadd_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), 2.0, 6.0)); __m128d test_mm_maskz_fnmadd_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { // CHECK-LABEL: test_mm_maskz_fnmadd_pd @@ -2917,6 +2939,30 @@ __m128d test_mm_maskz_fnmadd_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d TEST_CONSTEXPR(match_m128d(_mm_maskz_fnmadd_pd((__mmask8)0b10, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 0.0, -2.0)); TEST_CONSTEXPR(match_m128d(_mm_maskz_fnmadd_pd((__mmask8)0b01, (__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 2.0, 0.0)); +__m128d test_mm_mask_fnmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { + // CHECK-LABEL: test_mm_mask_fnmsub_pd + // CHECK: fneg <2 x double> %{{.*}} + // CHECK: fneg <2 x double> %{{.*}} + // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> + // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} + return _mm_mask_fnmsub_pd(__A, __U, __B, __C); +} +TEST_CONSTEXPR(match_m128d(_mm_mask_fnmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, -14.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask_fnmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), -8.0, 2.0)); + +__m128d test_mm_mask3_fnmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fnmsub_pd + // CHECK: fneg <2 x double> %{{.*}} + // CHECK: fneg <2 x double> %{{.*}} + // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> + // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} + return _mm_mask3_fnmsub_pd(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, -14.0)); +TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), -8.0, 6.0)); + __m128d test_mm_maskz_fnmsub_pd(__mmask8 __U, __m128d __A, __m128d __B, __m128d __C) { // CHECK-LABEL: test_mm_maskz_fnmsub_pd // CHECK: fneg <2 x double> %{{.*}} @@ -2939,6 +2985,26 @@ __m256d test_mm256_mask_fmadd_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d TEST_CONSTEXPR(match_m256d(_mm256_mask_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, 32.0, 44.0)); TEST_CONSTEXPR(match_m256d(_mm256_mask_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 14.0, 22.0, 3.0, 4.0)); +__m256d test_mm256_mask3_fmadd_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fmadd_pd + // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} + return _mm256_mask3_fmadd_pd(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, 32.0, 44.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), 14.0, 22.0, 11.0, 12.0)); + +__m256d test_mm256_maskz_fmadd_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { + // CHECK-LABEL: test_mm256_maskz_fmadd_pd + // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} + return _mm256_maskz_fmadd_pd(__U, __A, __B, __C); +} +TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmadd_pd((__mmask8)0b1100, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 0.0, 0.0, 32.0, 44.0)); +TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmadd_pd((__mmask8)0b0011, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 14.0, 22.0, 0.0, 0.0)); + __m256d test_mm256_mask_fmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { // CHECK-LABEL: test_mm256_mask_fmsub_pd // CHECK: fneg <4 x double> %{{.*}} @@ -2950,47 +3016,49 @@ __m256d test_mm256_mask_fmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d TEST_CONSTEXPR(match_m256d(_mm256_mask_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, 10.0, 20.0)); TEST_CONSTEXPR(match_m256d(_mm256_mask_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), -4.0, 2.0, 3.0, 4.0)); -__m256d test_mm256_mask3_fmadd_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fmadd_pd +__m256d test_mm256_mask3_fmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fmsub_pd + // CHECK: fneg <4 x double> %{{.*}} // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask3_fmadd_pd(__A, __B, __C, __U); + return _mm256_mask3_fmsub_pd(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, 32.0, 44.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), 14.0, 22.0, 11.0, 12.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, 10.0, 20.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), -4.0, 2.0, 11.0, 12.0)); -__m256d test_mm256_mask3_fnmadd_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fnmadd_pd +__m256d test_mm256_maskz_fmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { + // CHECK-LABEL: test_mm256_maskz_fmsub_pd // CHECK: fneg <4 x double> %{{.*}} // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask3_fnmadd_pd(__A, __B, __C, __U); + return _mm256_maskz_fmsub_pd(__U, __A, __B, __C); } -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, -10.0, -20.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), 4.0, -2.0, 11.0, 12.0)); +TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmsub_pd((__mmask8)0b1100, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 0.0, 0.0, 10.0, 20.0)); +TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmsub_pd((__mmask8)0b0011, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), -4.0, 2.0, 0.0, 0.0)); -__m256d test_mm256_maskz_fmadd_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { - // CHECK-LABEL: test_mm256_maskz_fmadd_pd +__m256d test_mm256_mask_fnmadd_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { + // CHECK-LABEL: test_mm256_mask_fnmadd_pd + // CHECK: fneg <4 x double> %{{.*}} // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_maskz_fmadd_pd(__U, __A, __B, __C); + return _mm256_mask_fnmadd_pd(__A, __U, __B, __C); } -TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmadd_pd((__mmask8)0b1100, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 0.0, 0.0, 32.0, 44.0)); -TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmadd_pd((__mmask8)0b0011, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 14.0, 22.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, -10.0, -20.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 4.0, -2.0, 3.0, 4.0)); -__m256d test_mm256_maskz_fmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { - // CHECK-LABEL: test_mm256_maskz_fmsub_pd +__m256d test_mm256_mask3_fnmadd_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fnmadd_pd // CHECK: fneg <4 x double> %{{.*}} // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_maskz_fmsub_pd(__U, __A, __B, __C); + return _mm256_mask3_fnmadd_pd(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmsub_pd((__mmask8)0b1100, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 0.0, 0.0, 10.0, 20.0)); -TEST_CONSTEXPR(match_m256d(_mm256_maskz_fmsub_pd((__mmask8)0b0011, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), -4.0, 2.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, -10.0, -20.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), 4.0, -2.0, 11.0, 12.0)); __m256d test_mm256_maskz_fnmadd_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { // CHECK-LABEL: test_mm256_maskz_fnmadd_pd @@ -3003,6 +3071,30 @@ __m256d test_mm256_maskz_fnmadd_pd(__mmask8 __U, __m256d __A, __m256d __B, __m25 TEST_CONSTEXPR(match_m256d(_mm256_maskz_fnmadd_pd((__mmask8)0b1100, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 0.0, 0.0, -10.0, -20.0)); TEST_CONSTEXPR(match_m256d(_mm256_maskz_fnmadd_pd((__mmask8)0b0011, (__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 4.0, -2.0, 0.0, 0.0)); +__m256d test_mm256_mask_fnmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { + // CHECK-LABEL: test_mm256_mask_fnmsub_pd + // CHECK: fneg <4 x double> %{{.*}} + // CHECK: fneg <4 x double> %{{.*}} + // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} + return _mm256_mask_fnmsub_pd(__A, __U, __B, __C); +} +TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, -32.0, -44.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), -14.0, -22.0, 3.0, 4.0)); + +__m256d test_mm256_mask3_fnmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fnmsub_pd + // CHECK: fneg <4 x double> %{{.*}} + // CHECK: fneg <4 x double> %{{.*}} + // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} + return _mm256_mask3_fnmsub_pd(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, -32.0, -44.0)); +TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), -14.0, -22.0, 11.0, 12.0)); + __m256d test_mm256_maskz_fnmsub_pd(__mmask8 __U, __m256d __A, __m256d __B, __m256d __C) { // CHECK-LABEL: test_mm256_maskz_fnmsub_pd // CHECK: fneg <4 x double> %{{.*}} @@ -3025,6 +3117,26 @@ __m128 test_mm_mask_fmadd_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { TEST_CONSTEXPR(match_m128(_mm_mask_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, 32.0, 44.0)); TEST_CONSTEXPR(match_m128(_mm_mask_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 14.0, 22.0, 3.0, 4.0)); +__m128 test_mm_mask3_fmadd_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fmadd_ps + // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} + return _mm_mask3_fmadd_ps(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m128(_mm_mask3_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, 32.0, 44.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), 14.0, 22.0, 11.0, 12.0)); + +__m128 test_mm_maskz_fmadd_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { + // CHECK-LABEL: test_mm_maskz_fmadd_ps + // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} + return _mm_maskz_fmadd_ps(__U, __A, __B, __C); +} +TEST_CONSTEXPR(match_m128(_mm_maskz_fmadd_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, 32.0, 44.0)); +TEST_CONSTEXPR(match_m128(_mm_maskz_fmadd_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 14.0, 22.0, 0.0, 0.0)); + __m128 test_mm_mask_fmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { // CHECK-LABEL: test_mm_mask_fmsub_ps // CHECK: fneg <4 x float> %{{.*}} @@ -3036,47 +3148,49 @@ __m128 test_mm_mask_fmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { TEST_CONSTEXPR(match_m128(_mm_mask_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, 10.0, 20.0)); TEST_CONSTEXPR(match_m128(_mm_mask_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -4.0, 2.0, 3.0, 4.0)); -__m128 test_mm_mask3_fmadd_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fmadd_ps +__m128 test_mm_mask3_fmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fmsub_ps + // CHECK: fneg <4 x float> %{{.*}} // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask3_fmadd_ps(__A, __B, __C, __U); + return _mm_mask3_fmsub_ps(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m128(_mm_mask3_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, 32.0, 44.0)); -TEST_CONSTEXPR(match_m128(_mm_mask3_fmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), 14.0, 22.0, 11.0, 12.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, 10.0, 20.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), -4.0, 2.0, 11.0, 12.0)); -__m128 test_mm_mask3_fnmadd_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fnmadd_ps +__m128 test_mm_maskz_fmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { + // CHECK-LABEL: test_mm_maskz_fmsub_ps // CHECK: fneg <4 x float> %{{.*}} // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask3_fnmadd_ps(__A, __B, __C, __U); + return _mm_maskz_fmsub_ps(__U, __A, __B, __C); } -TEST_CONSTEXPR(match_m128(_mm_mask3_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, -10.0, -20.0)); -TEST_CONSTEXPR(match_m128(_mm_mask3_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), 4.0, -2.0, 11.0, 12.0)); +TEST_CONSTEXPR(match_m128(_mm_maskz_fmsub_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, 10.0, 20.0)); +TEST_CONSTEXPR(match_m128(_mm_maskz_fmsub_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -4.0, 2.0, 0.0, 0.0)); -__m128 test_mm_maskz_fmadd_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { - // CHECK-LABEL: test_mm_maskz_fmadd_ps +__m128 test_mm_mask_fnmadd_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { + // CHECK-LABEL: test_mm_mask_fnmadd_ps + // CHECK: fneg <4 x float> %{{.*}} // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_maskz_fmadd_ps(__U, __A, __B, __C); + return _mm_mask_fnmadd_ps(__A, __U, __B, __C); } -TEST_CONSTEXPR(match_m128(_mm_maskz_fmadd_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, 32.0, 44.0)); -TEST_CONSTEXPR(match_m128(_mm_maskz_fmadd_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 14.0, 22.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m128(_mm_mask_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, -10.0, -20.0)); +TEST_CONSTEXPR(match_m128(_mm_mask_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 4.0, -2.0, 3.0, 4.0)); -__m128 test_mm_maskz_fmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { - // CHECK-LABEL: test_mm_maskz_fmsub_ps +__m128 test_mm_mask3_fnmadd_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fnmadd_ps // CHECK: fneg <4 x float> %{{.*}} // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_maskz_fmsub_ps(__U, __A, __B, __C); + return _mm_mask3_fnmadd_ps(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m128(_mm_maskz_fmsub_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, 10.0, 20.0)); -TEST_CONSTEXPR(match_m128(_mm_maskz_fmsub_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -4.0, 2.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, -10.0, -20.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), 4.0, -2.0, 11.0, 12.0)); __m128 test_mm_maskz_fnmadd_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { // CHECK-LABEL: test_mm_maskz_fnmadd_ps @@ -3089,6 +3203,30 @@ __m128 test_mm_maskz_fnmadd_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) TEST_CONSTEXPR(match_m128(_mm_maskz_fnmadd_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, -10.0, -20.0)); TEST_CONSTEXPR(match_m128(_mm_maskz_fnmadd_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 4.0, -2.0, 0.0, 0.0)); +__m128 test_mm_mask_fnmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { + // CHECK-LABEL: test_mm_mask_fnmsub_ps + // CHECK: fneg <4 x float> %{{.*}} + // CHECK: fneg <4 x float> %{{.*}} + // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} + return _mm_mask_fnmsub_ps(__A, __U, __B, __C); +} +TEST_CONSTEXPR(match_m128(_mm_mask_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, -32.0, -44.0)); +TEST_CONSTEXPR(match_m128(_mm_mask_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -14.0, -22.0, 3.0, 4.0)); + +__m128 test_mm_mask3_fnmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm_mask3_fnmsub_ps + // CHECK: fneg <4 x float> %{{.*}} + // CHECK: fneg <4 x float> %{{.*}} + // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> + // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} + return _mm_mask3_fnmsub_ps(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m128(_mm_mask3_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, -32.0, -44.0)); +TEST_CONSTEXPR(match_m128(_mm_mask3_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), -14.0, -22.0, 11.0, 12.0)); + __m128 test_mm_maskz_fnmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) { // CHECK-LABEL: test_mm_maskz_fnmsub_ps // CHECK: fneg <4 x float> %{{.*}} @@ -3101,14 +3239,32 @@ __m128 test_mm_maskz_fnmsub_ps(__mmask8 __U, __m128 __A, __m128 __B, __m128 __C) TEST_CONSTEXPR(match_m128(_mm_maskz_fnmsub_ps((__mmask8)0b1100, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 0.0, 0.0, -32.0, -44.0)); TEST_CONSTEXPR(match_m128(_mm_maskz_fnmsub_ps((__mmask8)0b0011, (__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -14.0, -22.0, 0.0, 0.0)); -__m256 test_mm256_mask_fmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { - // CHECK-LABEL: test_mm256_mask_fmadd_ps +__m256 test_mm256_mask_fmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { + // CHECK-LABEL: test_mm256_mask_fmadd_ps + // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) + // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} + return _mm256_mask_fmadd_ps(__A, __U, __B, __C); +} +TEST_CONSTEXPR(match_m256(_mm256_mask_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, 86.0, 106.0, 128.0, 152.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 26.0, 38.0, 52.0, 68.0, 5.0, 6.0, 7.0, 8.0)); + +__m256 test_mm256_mask3_fmadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fmadd_ps + // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) + // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} + return _mm256_mask3_fmadd_ps(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m256(_mm256_mask3_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, 86.0, 106.0, 128.0, 152.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask3_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), 26.0, 38.0, 52.0, 68.0, 21.0, 22.0, 23.0, 24.0)); + +__m256 test_mm256_maskz_fmadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { + // CHECK-LABEL: test_mm256_maskz_fmadd_ps // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask_fmadd_ps(__A, __U, __B, __C); + return _mm256_maskz_fmadd_ps(__U, __A, __B, __C); } -TEST_CONSTEXPR(match_m256(_mm256_mask_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, 86.0, 106.0, 128.0, 152.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 26.0, 38.0, 52.0, 68.0, 5.0, 6.0, 7.0, 8.0)); +TEST_CONSTEXPR(match_m256(_mm256_maskz_fmadd_ps((__mmask8)0b11110000, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 0.0, 0.0, 0.0, 0.0, 86.0, 106.0, 128.0, 152.0)); +TEST_CONSTEXPR(match_m256(_mm256_maskz_fmadd_ps((__mmask8)0b00001111, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 26.0, 38.0, 52.0, 68.0, 0.0, 0.0, 0.0, 0.0)); __m256 test_mm256_mask_fmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { // CHECK-LABEL: test_mm256_mask_fmsub_ps @@ -3120,43 +3276,45 @@ __m256 test_mm256_mask_fmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C TEST_CONSTEXPR(match_m256(_mm256_mask_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, 44.0, 62.0, 82.0, 104.0)); TEST_CONSTEXPR(match_m256(_mm256_mask_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), -8.0, 2.0, 14.0, 28.0, 5.0, 6.0, 7.0, 8.0)); -__m256 test_mm256_mask3_fmadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fmadd_ps +__m256 test_mm256_mask3_fmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fmsub_ps + // CHECK: fneg <8 x float> %{{.*}} // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask3_fmadd_ps(__A, __B, __C, __U); + return _mm256_mask3_fmsub_ps(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m256(_mm256_mask3_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, 86.0, 106.0, 128.0, 152.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask3_fmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), 26.0, 38.0, 52.0, 68.0, 21.0, 22.0, 23.0, 24.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask3_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, 44.0, 62.0, 82.0, 104.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask3_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), -8.0, 2.0, 14.0, 28.0, 21.0, 22.0, 23.0, 24.0)); -__m256 test_mm256_mask3_fnmadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fnmadd_ps +__m256 test_mm256_maskz_fmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { + // CHECK-LABEL: test_mm256_maskz_fmsub_ps // CHECK: fneg <8 x float> %{{.*}} // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask3_fnmadd_ps(__A, __B, __C, __U); + return _mm256_maskz_fmsub_ps(__U, __A, __B, __C); } -TEST_CONSTEXPR(match_m256( _mm256_mask3_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, -44.0, -62.0, -82.0, -104.0)); -TEST_CONSTEXPR(match_m256( _mm256_mask3_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), 8.0, -2.0, -14.0, -28.0, 21.0, 22.0, 23.0, 24.0)); +TEST_CONSTEXPR(match_m256(_mm256_maskz_fmsub_ps((__mmask8)0b11110000, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 0.0, 0.0, 0.0, 0.0, 44.0, 62.0, 82.0, 104.0)); +TEST_CONSTEXPR(match_m256(_mm256_maskz_fmsub_ps((__mmask8)0b00001111, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), -8.0, 2.0, 14.0, 28.0, 0.0, 0.0, 0.0, 0.0)); -__m256 test_mm256_maskz_fmadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { - // CHECK-LABEL: test_mm256_maskz_fmadd_ps +__m256 test_mm256_mask_fnmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { + // CHECK-LABEL: test_mm256_mask_fnmadd_ps + // CHECK: fneg <8 x float> %{{.*}} // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_maskz_fmadd_ps(__U, __A, __B, __C); + return _mm256_mask_fnmadd_ps(__A, __U, __B, __C); } -TEST_CONSTEXPR(match_m256(_mm256_maskz_fmadd_ps((__mmask8)0b11110000, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 0.0, 0.0, 0.0, 0.0, 86.0, 106.0, 128.0, 152.0)); -TEST_CONSTEXPR(match_m256(_mm256_maskz_fmadd_ps((__mmask8)0b00001111, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 26.0, 38.0, 52.0, 68.0, 0.0, 0.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, -44.0, -62.0, -82.0, -104.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 8.0, -2.0, -14.0, -28.0, 5.0, 6.0, 7.0, 8.0)); -__m256 test_mm256_maskz_fmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { - // CHECK-LABEL: test_mm256_maskz_fmsub_ps +__m256 test_mm256_mask3_fnmadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fnmadd_ps // CHECK: fneg <8 x float> %{{.*}} // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_maskz_fmsub_ps(__U, __A, __B, __C); + return _mm256_mask3_fnmadd_ps(__A, __B, __C, __U); } -TEST_CONSTEXPR(match_m256(_mm256_maskz_fmsub_ps((__mmask8)0b11110000, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 0.0, 0.0, 0.0, 0.0, 44.0, 62.0, 82.0, 104.0)); -TEST_CONSTEXPR(match_m256(_mm256_maskz_fmsub_ps((__mmask8)0b00001111, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), -8.0, 2.0, 14.0, 28.0, 0.0, 0.0, 0.0, 0.0)); +TEST_CONSTEXPR(match_m256( _mm256_mask3_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, -44.0, -62.0, -82.0, -104.0)); +TEST_CONSTEXPR(match_m256( _mm256_mask3_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), 8.0, -2.0, -14.0, -28.0, 21.0, 22.0, 23.0, 24.0)); __m256 test_mm256_maskz_fnmadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { // CHECK-LABEL: test_mm256_maskz_fnmadd_ps @@ -3168,6 +3326,28 @@ __m256 test_mm256_maskz_fnmadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 _ TEST_CONSTEXPR(match_m256(_mm256_maskz_fnmadd_ps((__mmask8)0b11110000, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 0.0, 0.0, 0.0, 0.0, -44.0, -62.0, -82.0, -104.0)); TEST_CONSTEXPR(match_m256(_mm256_maskz_fnmadd_ps((__mmask8)0b00001111, (__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 8.0, -2.0, -14.0, -28.0, 0.0, 0.0, 0.0, 0.0)); +__m256 test_mm256_mask_fnmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { + // CHECK-LABEL: test_mm256_mask_fnmsub_ps + // CHECK: fneg <8 x float> %{{.*}} + // CHECK: fneg <8 x float> %{{.*}} + // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) + // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} + return _mm256_mask_fnmsub_ps(__A, __U, __B, __C); +} +TEST_CONSTEXPR(match_m256(_mm256_mask_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, -86.0, -106.0, -128.0, -152.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), -26.0, -38.0, -52.0, -68.0, 5.0, 6.0, 7.0, 8.0)); + +__m256 test_mm256_mask3_fnmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { + // CHECK-LABEL: test_mm256_mask3_fnmsub_ps + // CHECK: fneg <8 x float> %{{.*}} + // CHECK: fneg <8 x float> %{{.*}} + // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) + // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} + return _mm256_mask3_fnmsub_ps(__A, __B, __C, __U); +} +TEST_CONSTEXPR(match_m256(_mm256_mask3_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, -86.0, -106.0, -128.0, -152.0)); +TEST_CONSTEXPR(match_m256(_mm256_mask3_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), -26.0, -38.0, -52.0, -68.0, 21.0, 22.0, 23.0, 24.0)); + __m256 test_mm256_maskz_fnmsub_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 __C) { // CHECK-LABEL: test_mm256_maskz_fnmsub_ps // CHECK: fneg <8 x float> %{{.*}} @@ -3354,49 +3534,6 @@ __m256 test_mm256_maskz_fmsubadd_ps(__mmask8 __U, __m256 __A, __m256 __B, __m256 return _mm256_maskz_fmsubadd_ps(__U, __A, __B, __C); } -__m128d test_mm_mask3_fmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fmsub_pd - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> - // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask3_fmsub_pd(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m128d(_mm_mask3_fmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, 2.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask3_fmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), -2.0, 6.0)); - -__m256d test_mm256_mask3_fmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fmsub_pd - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask3_fmsub_pd(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, 10.0, 20.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), -4.0, 2.0, 11.0, 12.0)); - -__m128 test_mm_mask3_fmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fmsub_ps - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask3_fmsub_ps(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m128(_mm_mask3_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, 10.0, 20.0)); -TEST_CONSTEXPR(match_m128(_mm_mask3_fmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), -4.0, 2.0, 11.0, 12.0)); - -__m256 test_mm256_mask3_fmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fmsub_ps - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) - // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask3_fmsub_ps(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m256(_mm256_mask3_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, 44.0, 62.0, 82.0, 104.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask3_fmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), -8.0, 2.0, 14.0, 28.0, 21.0, 22.0, 23.0, 24.0)); - __m128d test_mm_mask3_fmsubadd_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { // CHECK-LABEL: test_mm_mask3_fmsubadd_pd // CHECK: [[NEG:%.+]] = fneg <2 x double> %{{.+}} @@ -3432,143 +3569,6 @@ __m256 test_mm256_mask3_fmsubadd_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 return _mm256_mask3_fmsubadd_ps(__A, __B, __C, __U); } -__m128d test_mm_mask_fnmadd_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { - // CHECK-LABEL: test_mm_mask_fnmadd_pd - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> - // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask_fnmadd_pd(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m128d(_mm_mask_fnmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, -2.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask_fnmadd_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 2.0, 2.0)); - -__m256d test_mm256_mask_fnmadd_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { - // CHECK-LABEL: test_mm256_mask_fnmadd_pd - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask_fnmadd_pd(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, -10.0, -20.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmadd_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 4.0, -2.0, 3.0, 4.0)); - -__m128 test_mm_mask_fnmadd_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { - // CHECK-LABEL: test_mm_mask_fnmadd_ps - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask_fnmadd_ps(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m128(_mm_mask_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, -10.0, -20.0)); -TEST_CONSTEXPR(match_m128(_mm_mask_fnmadd_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 4.0, -2.0, 3.0, 4.0)); - -__m256 test_mm256_mask_fnmadd_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { - // CHECK-LABEL: test_mm256_mask_fnmadd_ps - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) - // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask_fnmadd_ps(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m256(_mm256_mask_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, -44.0, -62.0, -82.0, -104.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask_fnmadd_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 8.0, -2.0, -14.0, -28.0, 5.0, 6.0, 7.0, 8.0)); - -__m128d test_mm_mask_fnmsub_pd(__m128d __A, __mmask8 __U, __m128d __B, __m128d __C) { - // CHECK-LABEL: test_mm_mask_fnmsub_pd - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> - // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask_fnmsub_pd(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m128d(_mm_mask_fnmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b10, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), 1.0, -14.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask_fnmsub_pd((__m128d){1.0, 2.0}, (__mmask8)0b01, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}), -8.0, 2.0)); - -__m128d test_mm_mask3_fnmsub_pd(__m128d __A, __m128d __B, __m128d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fnmsub_pd - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: fneg <2 x double> %{{.*}} - // CHECK: call {{.*}}<2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <2 x i32> <i32 0, i32 1> - // CHECK: select <2 x i1> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}} - return _mm_mask3_fnmsub_pd(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b10), 5.0, -14.0)); -TEST_CONSTEXPR(match_m128d(_mm_mask3_fnmsub_pd((__m128d){1.0, 2.0}, (__m128d){3.0, 4.0}, (__m128d){5.0, 6.0}, (__mmask8)0b01), -8.0, 6.0)); - -__m256d test_mm256_mask_fnmsub_pd(__m256d __A, __mmask8 __U, __m256d __B, __m256d __C) { - // CHECK-LABEL: test_mm256_mask_fnmsub_pd - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask_fnmsub_pd(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b1100, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), 1.0, 2.0, -32.0, -44.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__mmask8)0b0011, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }), -14.0, -22.0, 3.0, 4.0)); - -__m256d test_mm256_mask3_fnmsub_pd(__m256d __A, __m256d __B, __m256d __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fnmsub_pd - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: fneg <4 x double> %{{.*}} - // CHECK: call {{.*}}<4 x double> @llvm.fma.v4f64(<4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}} - return _mm256_mask3_fnmsub_pd(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b1100), 9.0, 10.0, -32.0, -44.0)); -TEST_CONSTEXPR(match_m256d(_mm256_mask3_fnmsub_pd((__m256d){ 1.0, 2.0, 3.0, 4.0 }, (__m256d){ 5.0, 6.0, 7.0, 8.0 }, (__m256d){ 9.0, 10.0, 11.0, 12.0 }, (__mmask8)0b0011), -14.0, -22.0, 11.0, 12.0)); - -__m128 test_mm_mask_fnmsub_ps(__m128 __A, __mmask8 __U, __m128 __B, __m128 __C) { - // CHECK-LABEL: test_mm_mask_fnmsub_ps - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask_fnmsub_ps(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m128(_mm_mask_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b1100, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), 1.0, 2.0, -32.0, -44.0)); -TEST_CONSTEXPR(match_m128(_mm_mask_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__mmask8)0b0011, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}), -14.0, -22.0, 3.0, 4.0)); - -__m128 test_mm_mask3_fnmsub_ps(__m128 __A, __m128 __B, __m128 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm_mask3_fnmsub_ps - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: fneg <4 x float> %{{.*}} - // CHECK: call {{.*}}<4 x float> @llvm.fma.v4f32(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) - // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> - // CHECK: select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}} - return _mm_mask3_fnmsub_ps(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m128(_mm_mask3_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b1100), 9.0, 10.0, -32.0, -44.0)); -TEST_CONSTEXPR(match_m128(_mm_mask3_fnmsub_ps((__m128){1.0, 2.0, 3.0, 4.0}, (__m128){5.0, 6.0, 7.0, 8.0}, (__m128){9.0, 10.0, 11.0, 12.0}, (__mmask8)0b0011), -14.0, -22.0, 11.0, 12.0)); - -__m256 test_mm256_mask_fnmsub_ps(__m256 __A, __mmask8 __U, __m256 __B, __m256 __C) { - // CHECK-LABEL: test_mm256_mask_fnmsub_ps - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) - // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask_fnmsub_ps(__A, __U, __B, __C); -} -TEST_CONSTEXPR(match_m256(_mm256_mask_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b11110000, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), 1.0, 2.0, 3.0, 4.0, -86.0, -106.0, -128.0, -152.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__mmask8)0b00001111, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}), -26.0, -38.0, -52.0, -68.0, 5.0, 6.0, 7.0, 8.0)); - -__m256 test_mm256_mask3_fnmsub_ps(__m256 __A, __m256 __B, __m256 __C, __mmask8 __U) { - // CHECK-LABEL: test_mm256_mask3_fnmsub_ps - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: fneg <8 x float> %{{.*}} - // CHECK: call {{.*}}<8 x float> @llvm.fma.v8f32(<8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}) - // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}} - return _mm256_mask3_fnmsub_ps(__A, __B, __C, __U); -} -TEST_CONSTEXPR(match_m256(_mm256_mask3_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b11110000), 17.0, 18.0, 19.0, 20.0, -86.0, -106.0, -128.0, -152.0)); -TEST_CONSTEXPR(match_m256(_mm256_mask3_fnmsub_ps((__m256){1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0}, (__m256){9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0}, (__m256){17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0}, (__mmask8)0b00001111), -26.0, -38.0, -52.0, -68.0, 21.0, 22.0, 23.0, 24.0)); - __m128d test_mm_mask_add_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) { // CHECK-LABEL: test_mm_mask_add_pd // CHECK: fadd <2 x double> %{{.*}}, %{{.*}} _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
