diff options
author | Ramy Elgammal <ramy.elgammal@arm.com> | 2022-10-05 17:05:20 +0100 |
---|---|---|
committer | Ramy Elgammal <ramy.elgammal@arm.com> | 2022-10-12 09:08:35 +0000 |
commit | c8cc024603cb1db084227196a52e562bf251d339 (patch) | |
tree | 1ceb20a44f67749d259626897432548963c698f1 /src/cpu/operators | |
parent | a9acd5f2218a9abebe8290d2a4861808e36805e8 (diff) | |
download | ComputeLibrary-c8cc024603cb1db084227196a52e562bf251d339.tar.gz |
Adding documentation section explaining how BF16 is used
Resolves: COMPMID-5494
Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com>
Change-Id: I8f512745855b8ca21181a9ab21323bfff6aeb866
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/c/VisualCompute/ComputeLibrary/+/458884
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: bsgcomp <bsgcomp@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8391
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/operators')
-rw-r--r-- | src/cpu/operators/CpuGemm.h | 2 | ||||
-rw-r--r-- | src/cpu/operators/internal/CpuGemmAssemblyDispatch.cpp | 6 |
2 files changed, 4 insertions, 4 deletions
diff --git a/src/cpu/operators/CpuGemm.h b/src/cpu/operators/CpuGemm.h index 8d34b22437..031f02b3fd 100644 --- a/src/cpu/operators/CpuGemm.h +++ b/src/cpu/operators/CpuGemm.h @@ -76,7 +76,7 @@ public: * |:------------|:-----------|:---------|:--------------| * |F32 |F32 |F32 |F32 | * |F16 |F16 |F16 |F16 | - * |BFLOAT16 |BFLOAT16 |BFLOAT16 |BFLOAT16 | + * |BFLOAT16 |BFLOAT16 |BFLOAT16 |FP32 | * * @note GEMM: General Matrix Multiply - [alpha * A * B + beta * C]. * @note GEMM: The tensors a, b, c, d must have the same data type. You should not mix data types when calling this function. diff --git a/src/cpu/operators/internal/CpuGemmAssemblyDispatch.cpp b/src/cpu/operators/internal/CpuGemmAssemblyDispatch.cpp index 77da83070b..ab668681ad 100644 --- a/src/cpu/operators/internal/CpuGemmAssemblyDispatch.cpp +++ b/src/cpu/operators/internal/CpuGemmAssemblyDispatch.cpp @@ -716,7 +716,7 @@ Status CpuGemmAssemblyDispatch::has_opt_impl(arm_compute::WeightFormat &expected if(d->data_type() == DataType::S32) { ARM_COMPUTE_RETURN_ERROR_ON_MSG(!(arm_gemm::has_opt_gemm<uint8_t, uint32_t, arm_gemm::Nothing>(arm_gemm_expected_wf, args, {})), - "We could not find an optimized kernel for U8/QASYMM8 input and S32 output"); + "We could not find an optimized kernel for U8/QASYMM8 input and U32 output"); } else { @@ -734,7 +734,7 @@ Status CpuGemmAssemblyDispatch::has_opt_impl(arm_compute::WeightFormat &expected else { ARM_COMPUTE_RETURN_ERROR_ON_MSG(!(arm_gemm::has_opt_gemm<int8_t, int8_t, arm_gemm::Requantize32>(arm_gemm_expected_wf, args, {})), - "We could not find an optimized kernel for S8 input and S32 output"); + "We could not find an optimized kernel for S8 input and S8 output"); } break; #endif /* __aarch64__ */ @@ -749,7 +749,7 @@ Status CpuGemmAssemblyDispatch::has_opt_impl(arm_compute::WeightFormat &expected #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC case DataType::F16: ARM_COMPUTE_RETURN_ERROR_ON_MSG(!(arm_gemm::has_opt_gemm<float16_t, float16_t, arm_gemm::Nothing>(arm_gemm_expected_wf, args, {})), - "We could not find an optimized kernel for BFLOAT16 input and F32 output"); + "We could not find an optimized kernel for F16 input and F16 output"); break; #endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ default: |