diff options
| author | David Mansell <David.Mansell@arm.com> | 2024-05-03 13:24:48 +0100 |
|---|---|---|
| committer | Pablo Marquez Tello <pablo.tello@arm.com> | 2024-05-03 13:33:02 +0000 |
| commit | c22e1263ba3a6945ceb1fdccb33eac512fd156fb (patch) | |
| tree | 0240a38789ef97e5f1ccbc2d049acbdec25379ef | |
| parent | 0c5ba9ee5ad36aeb6d684c20d473a27f482e83fb (diff) | |
| download | ComputeLibrary-c22e1263ba3a6945ceb1fdccb33eac512fd156fb.tar.gz | |
arm_gemm: fix SVE check on fast mode kernels.
SVE BF16 kernels need to check for svebf16(), not just bf16().
Change-Id: I89494aac40166eba59719bed9822194a48ac282d
Signed-off-by: David Mansell <David.Mansell@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11520
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
| -rw-r--r-- | src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp b/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp index f223dea59e..0c1d3a387b 100644 --- a/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp +++ b/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp @@ -199,14 +199,14 @@ GemmImplementation<float, float>::with_estimate( GemmImplementation<float, float>::with_estimate( GemmMethod::GEMM_HYBRID, "sve_hybrid_fp32bf16fp32_mmla_6x4VL", - [](const GemmArgs &args) { return args._fast_mode && args._ci->has_bf16(); }, + [](const GemmArgs &args) { return args._fast_mode && args._ci->has_svebf16(); }, [](const GemmArgs &args) { return GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_6x4VL, float, float>::estimate_cycles<float>(args); }, [](const GemmArgs &args) { return new GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_6x4VL, float, float>(args); } ), GemmImplementation<float, float>::with_estimate( GemmMethod::GEMM_HYBRID, "sve_hybrid_fp32bf16fp32_mmla_4x6VL", - [](const GemmArgs &args) { return args._fast_mode && args._ci->has_bf16(); }, + [](const GemmArgs &args) { return args._fast_mode && args._ci->has_svebf16(); }, [](const GemmArgs &args) { return GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_4x6VL, float, float>::estimate_cycles<float>(args); }, [](const GemmArgs &args) { return new GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_4x6VL, float, float>(args); } ), |