derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
194 lines
6.6 KiB
C++
194 lines
6.6 KiB
C++
/*===----------------- avxifmaintrin.h - IFMA intrinsics -------------------===
|
|
*
|
|
* Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
* See https://llvm.org/LICENSE.txt for license information.
|
|
* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
*
|
|
*===-----------------------------------------------------------------------===
|
|
*/
|
|
|
|
#ifndef __IMMINTRIN_H
|
|
#error "Never use <avxifmaintrin.h> directly; include <immintrin.h> instead."
|
|
#endif
|
|
|
|
#ifndef __AVXIFMAINTRIN_H
|
|
#define __AVXIFMAINTRIN_H
|
|
|
|
/* Define the default attributes for the functions in this file. */
|
|
#if defined(__cplusplus) && (__cplusplus >= 201103L)
|
|
#define __DEFAULT_FN_ATTRS128 \
|
|
__attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \
|
|
__min_vector_width__(128))) constexpr
|
|
#define __DEFAULT_FN_ATTRS256 \
|
|
__attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \
|
|
__min_vector_width__(256))) constexpr
|
|
#else
|
|
#define __DEFAULT_FN_ATTRS128 \
|
|
__attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \
|
|
__min_vector_width__(128)))
|
|
#define __DEFAULT_FN_ATTRS256 \
|
|
__attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \
|
|
__min_vector_width__(256)))
|
|
#endif
|
|
|
|
#if !defined(__AVX512IFMA__) && defined(__AVXIFMA__)
|
|
#define _mm_madd52hi_epu64(X, Y, Z) _mm_madd52hi_avx_epu64(X, Y, Z)
|
|
#define _mm_madd52lo_epu64(X, Y, Z) _mm_madd52lo_avx_epu64(X, Y, Z)
|
|
#define _mm256_madd52hi_epu64(X, Y, Z) _mm256_madd52hi_avx_epu64(X, Y, Z)
|
|
#define _mm256_madd52lo_epu64(X, Y, Z) _mm256_madd52lo_avx_epu64(X, Y, Z)
|
|
#endif
|
|
|
|
// must vex-encoding
|
|
|
|
/// Multiply packed unsigned 52-bit integers in each 64-bit element of \a __Y
|
|
/// and \a __Z to form a 104-bit intermediate result. Add the high 52-bit
|
|
/// unsigned integer from the intermediate result with the corresponding
|
|
/// unsigned 64-bit integer in \a __X, and store the results in \a dst.
|
|
///
|
|
/// \headerfile <immintrin.h>
|
|
///
|
|
/// \code
|
|
/// __m128i
|
|
/// _mm_madd52hi_avx_epu64 (__m128i __X, __m128i __Y, __m128i __Z)
|
|
/// \endcode
|
|
///
|
|
/// This intrinsic corresponds to the \c VPMADD52HUQ instruction.
|
|
///
|
|
/// \return
|
|
/// return __m128i dst.
|
|
/// \param __X
|
|
/// A 128-bit vector of [2 x i64]
|
|
/// \param __Y
|
|
/// A 128-bit vector of [2 x i64]
|
|
/// \param __Z
|
|
/// A 128-bit vector of [2 x i64]
|
|
///
|
|
/// \code{.operation}
|
|
/// FOR j := 0 to 1
|
|
/// i := j*64
|
|
/// tmp[127:0] := ZeroExtend64(__Y[i+51:i]) * ZeroExtend64(__Z[i+51:i])
|
|
/// dst[i+63:i] := __X[i+63:i] + ZeroExtend64(tmp[103:52])
|
|
/// ENDFOR
|
|
/// dst[MAX:128] := 0
|
|
/// \endcode
|
|
static __inline__ __m128i __DEFAULT_FN_ATTRS128
|
|
_mm_madd52hi_avx_epu64(__m128i __X, __m128i __Y, __m128i __Z) {
|
|
return (__m128i)__builtin_ia32_vpmadd52huq128((__v2di)__X, (__v2di)__Y,
|
|
(__v2di)__Z);
|
|
}
|
|
|
|
/// Multiply packed unsigned 52-bit integers in each 64-bit element of \a __Y
|
|
/// and \a __Z to form a 104-bit intermediate result. Add the high 52-bit
|
|
/// unsigned integer from the intermediate result with the corresponding
|
|
/// unsigned 64-bit integer in \a __X, and store the results in \a dst.
|
|
///
|
|
/// \headerfile <immintrin.h>
|
|
///
|
|
/// \code
|
|
/// __m256i
|
|
/// _mm256_madd52hi_avx_epu64 (__m256i __X, __m256i __Y, __m256i __Z)
|
|
/// \endcode
|
|
///
|
|
/// This intrinsic corresponds to the \c VPMADD52HUQ instruction.
|
|
///
|
|
/// \return
|
|
/// return __m256i dst.
|
|
/// \param __X
|
|
/// A 256-bit vector of [4 x i64]
|
|
/// \param __Y
|
|
/// A 256-bit vector of [4 x i64]
|
|
/// \param __Z
|
|
/// A 256-bit vector of [4 x i64]
|
|
///
|
|
/// \code{.operation}
|
|
/// FOR j := 0 to 3
|
|
/// i := j*64
|
|
/// tmp[127:0] := ZeroExtend64(__Y[i+51:i]) * ZeroExtend64(__Z[i+51:i])
|
|
/// dst[i+63:i] := __X[i+63:i] + ZeroExtend64(tmp[103:52])
|
|
/// ENDFOR
|
|
/// dst[MAX:256] := 0
|
|
/// \endcode
|
|
static __inline__ __m256i __DEFAULT_FN_ATTRS256
|
|
_mm256_madd52hi_avx_epu64(__m256i __X, __m256i __Y, __m256i __Z) {
|
|
return (__m256i)__builtin_ia32_vpmadd52huq256((__v4di)__X, (__v4di)__Y,
|
|
(__v4di)__Z);
|
|
}
|
|
|
|
/// Multiply packed unsigned 52-bit integers in each 64-bit element of \a __Y
|
|
/// and \a __Z to form a 104-bit intermediate result. Add the low 52-bit
|
|
/// unsigned integer from the intermediate result with the corresponding
|
|
/// unsigned 64-bit integer in \a __X, and store the results in \a dst.
|
|
///
|
|
/// \headerfile <immintrin.h>
|
|
///
|
|
/// \code
|
|
/// __m128i
|
|
/// _mm_madd52lo_avx_epu64 (__m128i __X, __m128i __Y, __m128i __Z)
|
|
/// \endcode
|
|
///
|
|
/// This intrinsic corresponds to the \c VPMADD52LUQ instruction.
|
|
///
|
|
/// \return
|
|
/// return __m128i dst.
|
|
/// \param __X
|
|
/// A 128-bit vector of [2 x i64]
|
|
/// \param __Y
|
|
/// A 128-bit vector of [2 x i64]
|
|
/// \param __Z
|
|
/// A 128-bit vector of [2 x i64]
|
|
///
|
|
/// \code{.operation}
|
|
/// FOR j := 0 to 1
|
|
/// i := j*64
|
|
/// tmp[127:0] := ZeroExtend64(__Y[i+51:i]) * ZeroExtend64(__Z[i+51:i])
|
|
/// dst[i+63:i] := __X[i+63:i] + ZeroExtend64(tmp[51:0])
|
|
/// ENDFOR
|
|
/// dst[MAX:128] := 0
|
|
/// \endcode
|
|
static __inline__ __m128i __DEFAULT_FN_ATTRS128
|
|
_mm_madd52lo_avx_epu64(__m128i __X, __m128i __Y, __m128i __Z) {
|
|
return (__m128i)__builtin_ia32_vpmadd52luq128((__v2di)__X, (__v2di)__Y,
|
|
(__v2di)__Z);
|
|
}
|
|
|
|
/// Multiply packed unsigned 52-bit integers in each 64-bit element of \a __Y
|
|
/// and \a __Z to form a 104-bit intermediate result. Add the low 52-bit
|
|
/// unsigned integer from the intermediate result with the corresponding
|
|
/// unsigned 64-bit integer in \a __X, and store the results in \a dst.
|
|
///
|
|
/// \headerfile <immintrin.h>
|
|
///
|
|
/// \code
|
|
/// __m256i
|
|
/// _mm256_madd52lo_avx_epu64 (__m256i __X, __m256i __Y, __m256i __Z)
|
|
/// \endcode
|
|
///
|
|
/// This intrinsic corresponds to the \c VPMADD52LUQ instruction.
|
|
///
|
|
/// \return
|
|
/// return __m256i dst.
|
|
/// \param __X
|
|
/// A 256-bit vector of [4 x i64]
|
|
/// \param __Y
|
|
/// A 256-bit vector of [4 x i64]
|
|
/// \param __Z
|
|
/// A 256-bit vector of [4 x i64]
|
|
///
|
|
/// \code{.operation}
|
|
/// FOR j := 0 to 3
|
|
/// i := j*64
|
|
/// tmp[127:0] := ZeroExtend64(__Y[i+51:i]) * ZeroExtend64(__Z[i+51:i])
|
|
/// dst[i+63:i] := __X[i+63:i] + ZeroExtend64(tmp[51:0])
|
|
/// ENDFOR
|
|
/// dst[MAX:256] := 0
|
|
/// \endcode
|
|
static __inline__ __m256i __DEFAULT_FN_ATTRS256
|
|
_mm256_madd52lo_avx_epu64(__m256i __X, __m256i __Y, __m256i __Z) {
|
|
return (__m256i)__builtin_ia32_vpmadd52luq256((__v4di)__X, (__v4di)__Y,
|
|
(__v4di)__Z);
|
|
}
|
|
#undef __DEFAULT_FN_ATTRS128
|
|
#undef __DEFAULT_FN_ATTRS256
|
|
|
|
#endif // __AVXIFMAINTRIN_H
|