[AArch64] Implement FP8 SVE intrinsics for fused multiply-add (#118126)
This patch adds the following intrinsics:
* 8-bit floating-point multiply-add long to half-precision (bottom).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat16_t svmlalb[_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat16_t svmlalb[_n_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long to half-precision (bottom,
indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat16_t svmlalb_lane[_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
* 8-bit floating-point multiply-add long to half-precision (top).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat16_t svmlalt[_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat16_t svmlalt[_n_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long to half-precision (top,
indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat16_t svmlalt_lane[_f16_mf8]_fpm(svfloat16_t zda, svmfloat8_t zn,
svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision
(bottom bottom).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlallbb[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat32_t svmlallbb[_n_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision
(bottom bottom, indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlallbb_lane[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t
zn, svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision
(bottom top).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlallbt[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat32_t svmlallbt[_n_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision
(bottom top, indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlallbt_lane[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t
zn, svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision (top
bottom).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlalltb[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat32_t svmlalltb[_n_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision (top
bottom, indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlalltb_lane[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t
zn, svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision (top
top).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlalltt[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
svmfloat8_t zm, fpm_t fpm);
svfloat32_t svmlalltt[_n_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t zn,
mfloat8_t zm, fpm_t fpm);
* 8-bit floating-point multiply-add long long to single-precision (top
top, indexed).
// Only if (__ARM_FEATURE_SVE2 && __ARM_FEATURE_FP8FMA) ||
__ARM_FEATURE_SSVE_FP8FMA
svfloat32_t svmlalltt_lane[_f32_mf8]_fpm(svfloat32_t zda, svmfloat8_t
zn, svmfloat8_t zm,
uint64_t imm0_15, fpm_t fpm);
Welcome to the LLVM project!
This repository contains the source code for LLVM, a toolkit for the construction of highly optimized compilers, optimizers, and run-time environments.
The LLVM project has multiple components. The core of the project is itself called “LLVM”. This contains all of the tools, libraries, and header files needed to process intermediate representations and convert them into object files. Tools include an assembler, disassembler, bitcode analyzer, and bitcode optimizer.
C-like languages use the Clang frontend. This component compiles C, C++, Objective-C, and Objective-C++ code into LLVM bitcode -- and from there into object files, using LLVM.
Other components include: the libc++ C++ standard library, the LLD linker, and more.
Consult the Getting Started with LLVM page for information on building and running LLVM.
For information on how to contribute to the LLVM project, please take a look at the Contributing to LLVM guide.
Join the LLVM Discourse forums, Discord chat, LLVM Office Hours or Regular sync-ups.
The LLVM project has adopted a code of conduct for participants to all modes of communication within the project.