From 3ad23c9b71f40da84595797a7130773d63880786 Mon Sep 17 00:00:00 2001 From: sayantn Date: Mon, 2 Mar 2026 00:32:24 +0530 Subject: [PATCH] Fix LLVM intrinsic signatures for AVX-VNNI --- crates/core_arch/src/x86/avx512vnni.rs | 160 ++++++++++++------------- 1 file changed, 80 insertions(+), 80 deletions(-) diff --git a/crates/core_arch/src/x86/avx512vnni.rs b/crates/core_arch/src/x86/avx512vnni.rs index 49b790b151..8cd8764f24 100644 --- a/crates/core_arch/src/x86/avx512vnni.rs +++ b/crates/core_arch/src/x86/avx512vnni.rs @@ -12,7 +12,7 @@ use stdarch_test::assert_instr; #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssd))] pub fn _mm512_dpwssd_epi32(src: __m512i, a: __m512i, b: __m512i) -> __m512i { - unsafe { transmute(vpdpwssd(src.as_i32x16(), a.as_i32x16(), b.as_i32x16())) } + unsafe { transmute(vpdpwssd(src.as_i32x16(), a.as_i16x32(), b.as_i16x32())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -51,7 +51,7 @@ pub fn _mm512_maskz_dpwssd_epi32(k: __mmask16, src: __m512i, a: __m512i, b: __m5 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssd))] pub fn _mm256_dpwssd_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwssd256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwssd256(src.as_i32x8(), a.as_i16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst. @@ -62,7 +62,7 @@ pub fn _mm256_dpwssd_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssd))] pub fn _mm256_dpwssd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwssd256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwssd256(src.as_i32x8(), a.as_i16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -101,7 +101,7 @@ pub fn _mm256_maskz_dpwssd_epi32(k: __mmask8, src: __m256i, a: __m256i, b: __m25 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssd))] pub fn _mm_dpwssd_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwssd128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwssd128(src.as_i32x4(), a.as_i16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst. @@ -112,7 +112,7 @@ pub fn _mm_dpwssd_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssd))] pub fn _mm_dpwssd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwssd128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwssd128(src.as_i32x4(), a.as_i16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -151,7 +151,7 @@ pub fn _mm_maskz_dpwssd_epi32(k: __mmask8, src: __m128i, a: __m128i, b: __m128i) #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssds))] pub fn _mm512_dpwssds_epi32(src: __m512i, a: __m512i, b: __m512i) -> __m512i { - unsafe { transmute(vpdpwssds(src.as_i32x16(), a.as_i32x16(), b.as_i32x16())) } + unsafe { transmute(vpdpwssds(src.as_i32x16(), a.as_i16x32(), b.as_i16x32())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -190,7 +190,7 @@ pub fn _mm512_maskz_dpwssds_epi32(k: __mmask16, src: __m512i, a: __m512i, b: __m #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssds))] pub fn _mm256_dpwssds_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwssds256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwssds256(src.as_i32x8(), a.as_i16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst. @@ -201,7 +201,7 @@ pub fn _mm256_dpwssds_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssds))] pub fn _mm256_dpwssds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwssds256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwssds256(src.as_i32x8(), a.as_i16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -240,7 +240,7 @@ pub fn _mm256_maskz_dpwssds_epi32(k: __mmask8, src: __m256i, a: __m256i, b: __m2 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssds))] pub fn _mm_dpwssds_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwssds128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwssds128(src.as_i32x4(), a.as_i16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst. @@ -251,7 +251,7 @@ pub fn _mm_dpwssds_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpwssds))] pub fn _mm_dpwssds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwssds128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwssds128(src.as_i32x4(), a.as_i16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -290,7 +290,7 @@ pub fn _mm_maskz_dpwssds_epi32(k: __mmask8, src: __m128i, a: __m128i, b: __m128i #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusd))] pub fn _mm512_dpbusd_epi32(src: __m512i, a: __m512i, b: __m512i) -> __m512i { - unsafe { transmute(vpdpbusd(src.as_i32x16(), a.as_i32x16(), b.as_i32x16())) } + unsafe { transmute(vpdpbusd(src.as_i32x16(), a.as_u8x64(), b.as_i8x64())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -329,7 +329,7 @@ pub fn _mm512_maskz_dpbusd_epi32(k: __mmask16, src: __m512i, a: __m512i, b: __m5 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusd))] pub fn _mm256_dpbusd_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbusd256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbusd256(src.as_i32x8(), a.as_u8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst. @@ -340,7 +340,7 @@ pub fn _mm256_dpbusd_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusd))] pub fn _mm256_dpbusd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbusd256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbusd256(src.as_i32x8(), a.as_u8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -379,7 +379,7 @@ pub fn _mm256_maskz_dpbusd_epi32(k: __mmask8, src: __m256i, a: __m256i, b: __m25 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusd))] pub fn _mm_dpbusd_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbusd128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbusd128(src.as_i32x4(), a.as_u8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst. @@ -390,7 +390,7 @@ pub fn _mm_dpbusd_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusd))] pub fn _mm_dpbusd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbusd128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbusd128(src.as_i32x4(), a.as_u8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -429,7 +429,7 @@ pub fn _mm_maskz_dpbusd_epi32(k: __mmask8, src: __m128i, a: __m128i, b: __m128i) #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusds))] pub fn _mm512_dpbusds_epi32(src: __m512i, a: __m512i, b: __m512i) -> __m512i { - unsafe { transmute(vpdpbusds(src.as_i32x16(), a.as_i32x16(), b.as_i32x16())) } + unsafe { transmute(vpdpbusds(src.as_i32x16(), a.as_u8x64(), b.as_i8x64())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -468,7 +468,7 @@ pub fn _mm512_maskz_dpbusds_epi32(k: __mmask16, src: __m512i, a: __m512i, b: __m #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusds))] pub fn _mm256_dpbusds_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbusds256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbusds256(src.as_i32x8(), a.as_u8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst. @@ -479,7 +479,7 @@ pub fn _mm256_dpbusds_avx_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusds))] pub fn _mm256_dpbusds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbusds256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbusds256(src.as_i32x8(), a.as_u8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -518,7 +518,7 @@ pub fn _mm256_maskz_dpbusds_epi32(k: __mmask8, src: __m256i, a: __m256i, b: __m2 #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusds))] pub fn _mm_dpbusds_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbusds128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbusds128(src.as_i32x4(), a.as_u8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst. @@ -529,7 +529,7 @@ pub fn _mm_dpbusds_avx_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[stable(feature = "stdarch_x86_avx512", since = "1.89")] #[cfg_attr(test, assert_instr(vpdpbusds))] pub fn _mm_dpbusds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbusds128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbusds128(src.as_i32x4(), a.as_u8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). @@ -570,7 +570,7 @@ pub fn _mm_maskz_dpbusds_epi32(k: __mmask8, src: __m128i, a: __m128i, b: __m128i #[cfg_attr(test, assert_instr(vpdpbssd))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbssd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbssd_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbssd_128(src.as_i32x4(), a.as_i8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit @@ -583,7 +583,7 @@ pub fn _mm_dpbssd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbssd))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbssd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbssd_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbssd_256(src.as_i32x8(), a.as_i8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit @@ -596,7 +596,7 @@ pub fn _mm256_dpbssd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpbssds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbssds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbssds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbssds_128(src.as_i32x4(), a.as_i8x16(), b.as_i8x16())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit @@ -609,7 +609,7 @@ pub fn _mm_dpbssds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbssds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbssds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbssds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbssds_256(src.as_i32x8(), a.as_i8x32(), b.as_i8x32())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit @@ -622,7 +622,7 @@ pub fn _mm256_dpbssds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpbsud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbsud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbsud_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbsud_128(src.as_i32x4(), a.as_i8x16(), b.as_u8x16())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit @@ -635,7 +635,7 @@ pub fn _mm_dpbsud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbsud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbsud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbsud_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbsud_256(src.as_i32x8(), a.as_i8x32(), b.as_u8x32())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit @@ -648,7 +648,7 @@ pub fn _mm256_dpbsud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpbsuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbsuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbsuds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbsuds_128(src.as_i32x4(), a.as_i8x16(), b.as_u8x16())) } } /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit @@ -661,7 +661,7 @@ pub fn _mm_dpbsuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbsuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbsuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbsuds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbsuds_256(src.as_i32x8(), a.as_i8x32(), b.as_u8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit @@ -674,7 +674,7 @@ pub fn _mm256_dpbsuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpbuud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbuud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbuud_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbuud_128(src.as_i32x4(), a.as_u8x16(), b.as_u8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit @@ -687,7 +687,7 @@ pub fn _mm_dpbuud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbuud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbuud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbuud_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbuud_256(src.as_i32x8(), a.as_u8x32(), b.as_u8x32())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit @@ -700,7 +700,7 @@ pub fn _mm256_dpbuud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpbuuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpbuuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpbuuds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpbuuds_128(src.as_i32x4(), a.as_u8x16(), b.as_u8x16())) } } /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit @@ -713,7 +713,7 @@ pub fn _mm_dpbuuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpbuuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpbuuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpbuuds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpbuuds_256(src.as_i32x8(), a.as_u8x32(), b.as_u8x32())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit @@ -726,7 +726,7 @@ pub fn _mm256_dpbuuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwsud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwsud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwsud_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwsud_128(src.as_i32x4(), a.as_i16x8(), b.as_u16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit @@ -739,7 +739,7 @@ pub fn _mm_dpwsud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwsud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwsud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwsud_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwsud_256(src.as_i32x8(), a.as_i16x16(), b.as_u16x16())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit @@ -752,7 +752,7 @@ pub fn _mm256_dpwsud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwsuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwsuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwsuds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwsuds_128(src.as_i32x4(), a.as_i16x8(), b.as_u16x8())) } } /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit @@ -765,7 +765,7 @@ pub fn _mm_dpwsuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwsuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwsuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwsuds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwsuds_256(src.as_i32x8(), a.as_i16x16(), b.as_u16x16())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit @@ -778,7 +778,7 @@ pub fn _mm256_dpwsuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwusd))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwusd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwusd_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwusd_128(src.as_i32x4(), a.as_u16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit @@ -791,7 +791,7 @@ pub fn _mm_dpwusd_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwusd))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwusd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwusd_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwusd_256(src.as_i32x8(), a.as_u16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit @@ -804,7 +804,7 @@ pub fn _mm256_dpwusd_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwusds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwusds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwusds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwusds_128(src.as_i32x4(), a.as_u16x8(), b.as_i16x8())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit @@ -817,7 +817,7 @@ pub fn _mm_dpwusds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwusds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwusds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwusds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwusds_256(src.as_i32x8(), a.as_u16x16(), b.as_i16x16())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit @@ -830,7 +830,7 @@ pub fn _mm256_dpwusds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwuud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwuud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwuud_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwuud_128(src.as_i32x4(), a.as_u16x8(), b.as_u16x8())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit @@ -843,7 +843,7 @@ pub fn _mm_dpwuud_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwuud))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwuud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwuud_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwuud_256(src.as_i32x8(), a.as_u16x16(), b.as_u16x16())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit @@ -856,7 +856,7 @@ pub fn _mm256_dpwuud_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { #[cfg_attr(test, assert_instr(vpdpwuuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm_dpwuuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { - unsafe { transmute(vpdpwuuds_128(src.as_i32x4(), a.as_i32x4(), b.as_i32x4())) } + unsafe { transmute(vpdpwuuds_128(src.as_i32x4(), a.as_u16x8(), b.as_u16x8())) } } /// Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit @@ -869,98 +869,98 @@ pub fn _mm_dpwuuds_epi32(src: __m128i, a: __m128i, b: __m128i) -> __m128i { #[cfg_attr(test, assert_instr(vpdpwuuds))] #[stable(feature = "stdarch_x86_avx512", since = "1.89")] pub fn _mm256_dpwuuds_epi32(src: __m256i, a: __m256i, b: __m256i) -> __m256i { - unsafe { transmute(vpdpwuuds_256(src.as_i32x8(), a.as_i32x8(), b.as_i32x8())) } + unsafe { transmute(vpdpwuuds_256(src.as_i32x8(), a.as_u16x16(), b.as_u16x16())) } } #[allow(improper_ctypes)] unsafe extern "C" { #[link_name = "llvm.x86.avx512.vpdpwssd.512"] - fn vpdpwssd(src: i32x16, a: i32x16, b: i32x16) -> i32x16; + fn vpdpwssd(src: i32x16, a: i16x32, b: i16x32) -> i32x16; #[link_name = "llvm.x86.avx512.vpdpwssd.256"] - fn vpdpwssd256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwssd256(src: i32x8, a: i16x16, b: i16x16) -> i32x8; #[link_name = "llvm.x86.avx512.vpdpwssd.128"] - fn vpdpwssd128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwssd128(src: i32x4, a: i16x8, b: i16x8) -> i32x4; #[link_name = "llvm.x86.avx512.vpdpwssds.512"] - fn vpdpwssds(src: i32x16, a: i32x16, b: i32x16) -> i32x16; + fn vpdpwssds(src: i32x16, a: i16x32, b: i16x32) -> i32x16; #[link_name = "llvm.x86.avx512.vpdpwssds.256"] - fn vpdpwssds256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwssds256(src: i32x8, a: i16x16, b: i16x16) -> i32x8; #[link_name = "llvm.x86.avx512.vpdpwssds.128"] - fn vpdpwssds128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwssds128(src: i32x4, a: i16x8, b: i16x8) -> i32x4; #[link_name = "llvm.x86.avx512.vpdpbusd.512"] - fn vpdpbusd(src: i32x16, a: i32x16, b: i32x16) -> i32x16; + fn vpdpbusd(src: i32x16, a: u8x64, b: i8x64) -> i32x16; #[link_name = "llvm.x86.avx512.vpdpbusd.256"] - fn vpdpbusd256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbusd256(src: i32x8, a: u8x32, b: i8x32) -> i32x8; #[link_name = "llvm.x86.avx512.vpdpbusd.128"] - fn vpdpbusd128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbusd128(src: i32x4, a: u8x16, b: i8x16) -> i32x4; #[link_name = "llvm.x86.avx512.vpdpbusds.512"] - fn vpdpbusds(src: i32x16, a: i32x16, b: i32x16) -> i32x16; + fn vpdpbusds(src: i32x16, a: u8x64, b: i8x64) -> i32x16; #[link_name = "llvm.x86.avx512.vpdpbusds.256"] - fn vpdpbusds256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbusds256(src: i32x8, a: u8x32, b: i8x32) -> i32x8; #[link_name = "llvm.x86.avx512.vpdpbusds.128"] - fn vpdpbusds128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbusds128(src: i32x4, a: u8x16, b: i8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbssd.128"] - fn vpdpbssd_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbssd_128(src: i32x4, a: i8x16, b: i8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbssd.256"] - fn vpdpbssd_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbssd_256(src: i32x8, a: i8x32, b: i8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpbssds.128"] - fn vpdpbssds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbssds_128(src: i32x4, a: i8x16, b: i8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbssds.256"] - fn vpdpbssds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbssds_256(src: i32x8, a: i8x32, b: i8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpbsud.128"] - fn vpdpbsud_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbsud_128(src: i32x4, a: i8x16, b: u8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbsud.256"] - fn vpdpbsud_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbsud_256(src: i32x8, a: i8x32, b: u8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpbsuds.128"] - fn vpdpbsuds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbsuds_128(src: i32x4, a: i8x16, b: u8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbsuds.256"] - fn vpdpbsuds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbsuds_256(src: i32x8, a: i8x32, b: u8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpbuud.128"] - fn vpdpbuud_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbuud_128(src: i32x4, a: u8x16, b: u8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbuud.256"] - fn vpdpbuud_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbuud_256(src: i32x8, a: u8x32, b: u8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpbuuds.128"] - fn vpdpbuuds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpbuuds_128(src: i32x4, a: u8x16, b: u8x16) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpbuuds.256"] - fn vpdpbuuds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpbuuds_256(src: i32x8, a: u8x32, b: u8x32) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwsud.128"] - fn vpdpwsud_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwsud_128(src: i32x4, a: i16x8, b: u16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwsud.256"] - fn vpdpwsud_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwsud_256(src: i32x8, a: i16x16, b: u16x16) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwsuds.128"] - fn vpdpwsuds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwsuds_128(src: i32x4, a: i16x8, b: u16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwsuds.256"] - fn vpdpwsuds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwsuds_256(src: i32x8, a: i16x16, b: u16x16) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwusd.128"] - fn vpdpwusd_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwusd_128(src: i32x4, a: u16x8, b: i16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwusd.256"] - fn vpdpwusd_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwusd_256(src: i32x8, a: u16x16, b: i16x16) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwusds.128"] - fn vpdpwusds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwusds_128(src: i32x4, a: u16x8, b: i16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwusds.256"] - fn vpdpwusds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwusds_256(src: i32x8, a: u16x16, b: i16x16) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwuud.128"] - fn vpdpwuud_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwuud_128(src: i32x4, a: u16x8, b: u16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwuud.256"] - fn vpdpwuud_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwuud_256(src: i32x8, a: u16x16, b: u16x16) -> i32x8; #[link_name = "llvm.x86.avx2.vpdpwuuds.128"] - fn vpdpwuuds_128(src: i32x4, a: i32x4, b: i32x4) -> i32x4; + fn vpdpwuuds_128(src: i32x4, a: u16x8, b: u16x8) -> i32x4; #[link_name = "llvm.x86.avx2.vpdpwuuds.256"] - fn vpdpwuuds_256(src: i32x8, a: i32x8, b: i32x8) -> i32x8; + fn vpdpwuuds_256(src: i32x8, a: u16x16, b: u16x16) -> i32x8; } #[cfg(test)]