diff options
| author | SiCong Li <sicong.li@arm.com> | 2023-01-09 15:34:20 +0000 |
|---|---|---|
| committer | SiCong Li <sicong.li@arm.com> | 2023-01-18 16:44:11 +0000 |
| commit | 1b6377b603c5833e52d704dc17db14f446ebc670 (patch) | |
| tree | cd4b03c6cc8aefe27b6b5f5fe87c559b26367d2d /arm_compute | |
| parent | a68df8dd8befdb53ec03412cf88113c13aa47c09 (diff) | |
| download | ComputeLibrary-1b6377b603c5833e52d704dc17db14f446ebc670.tar.gz | |
Add broadcast batched matmul validation cases
Related to: COMPMID-5660
Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: I2314c8b21acc638402c77080d59db2f3fed58fe2
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8911
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Mohmun02 <MohammedSuhail.Munshi@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute')
| -rw-r--r-- | arm_compute/runtime/CL/functions/CLGEMM.h | 5 | ||||
| -rw-r--r-- | arm_compute/runtime/NEON/functions/NEGEMM.h | 4 |
2 files changed, 7 insertions, 2 deletions
diff --git a/arm_compute/runtime/CL/functions/CLGEMM.h b/arm_compute/runtime/CL/functions/CLGEMM.h index 38a07ef9fb..b267bf1cf2 100644 --- a/arm_compute/runtime/CL/functions/CLGEMM.h +++ b/arm_compute/runtime/CL/functions/CLGEMM.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2021 Arm Limited. + * Copyright (c) 2016-2021, 2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -77,6 +77,9 @@ public: * * @note Whilst the first input tensor can be a vector, the second input tensor must be at least a matrix * + * @note Batched GEMM only allows RHS tensor's rank to be <= 3 + * @note Batched GEMM only supports broadcasting cases where RHS rank < LHS rank but not the other way around + * * @param[in] compile_context The compile context to be used. * @param[in] a First input tensor (Matrix or Vector A). Data types supported: F16/F32 * @param[in] b Second input tensor (Matrix B). Data type supported: same as @p a. diff --git a/arm_compute/runtime/NEON/functions/NEGEMM.h b/arm_compute/runtime/NEON/functions/NEGEMM.h index 7ce2521148..db15923165 100644 --- a/arm_compute/runtime/NEON/functions/NEGEMM.h +++ b/arm_compute/runtime/NEON/functions/NEGEMM.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2022 Arm Limited. + * Copyright (c) 2017-2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -66,6 +66,8 @@ public: * @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. * + * @note Batched GEMM only supports broadcasting cases where RHS rank < LHS rank but not the other way around + * * @param[in] a First input tensor (Matrix A or Vector A). Data type supported: BFLOAT16/F16/F32 * @param[in] b Second input tensor (Matrix B). Data type supported: same as @p a * @param[in] c Third input tensor (Matrix C). It can be a nullptr if just the multiplication between @p a and @p b is needed. Data type supported: same as @p a |