diff options
| author | Pablo Marquez Tello <pablo.tello@arm.com> | 2023-09-19 14:46:07 +0100 |
|---|---|---|
| committer | Pablo Marquez Tello <pablo.tello@arm.com> | 2023-09-21 10:20:28 +0000 |
| commit | f57d6ec5ff4305d2e388730f6dad004908e6e97a (patch) | |
| tree | 5efc93e5699e649057c57660b717b726cf607a7b /src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h | |
| parent | e071b5e31004b29afefaa96907032bfd2b4e5a43 (diff) | |
| download | ComputeLibrary-f57d6ec5ff4305d2e388730f6dad004908e6e97a.tar.gz | |
Gemm changes to enable fp16 in armv8a multi_isa builds
* Code guarded with __ARM_FEATURE_FP16_VECTOR_ARITHMETIC needs
to be moved to an fp16.cpp file to allow compilation with
-march=armv8.2-a+fp16
* fp16.cpp needs to use the templates vector_matrix_multiply_f16() and
matrix_matrix_multiply_f16 which had to be moved from impl.cpp to fp16.cpp
* Partially resolves MLCE-1102
Change-Id: Ic87440797d6f1653c815ab6565972206f5afd0ad
Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10345
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h')
| -rw-r--r-- | src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h | 9 |
1 files changed, 1 insertions, 8 deletions
diff --git a/src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h b/src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h index 6bf865a624..f9f1f247ac 100644 --- a/src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h +++ b/src/cpu/kernels/gemm_matrix_mul/generic/neon/impl.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2022 Arm Limited. + * Copyright (c) 2022-2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -30,13 +30,6 @@ namespace arm_compute { namespace cpu { -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -void vector_matrix_multiply_f16(const ITensor *lhs, const ITensor *rhs, ITensor *dst, const Window &window, const ThreadInfo &info, float alpha); - -void matrix_matrix_multiply_f16(const ITensor *lhs, const ITensor *rhs, ITensor *dst, const Window &window, const ThreadInfo &info, float alpha); - -#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC - void vector_matrix_multiply_f32(const ITensor *lhs, const ITensor *rhs, ITensor *dst, const Window &window, const ThreadInfo &info, float alpha); void matrix_matrix_multiply_f32(const ITensor *lhs, const ITensor *rhs, ITensor *dst, const Window &window, const ThreadInfo &info, float alpha); |