diff options
author | Ramy Elgammal <ramy.elgammal@arm.com> | 2023-03-09 21:15:37 +0000 |
---|---|---|
committer | Gunes Bayir <gunes.bayir@arm.com> | 2023-03-17 11:00:57 +0000 |
commit | 2b6ebfe4270b06b09e45f306e8384950aeca7e4e (patch) | |
tree | 44a65551aa32352a57090bf550b96cbf5b54222f /src/gpu/cl/kernels/ClNativeMatMulKernel.h | |
parent | 40aab11ff88ed432d5028478e531f8d0fc404d4c (diff) | |
download | ComputeLibrary-2b6ebfe4270b06b09e45f306e8384950aeca7e4e.tar.gz |
Implement OpenCL MatMul for Lhs NT Rhs T/NT FP32/16
- Implement ClNativeMatMulKernel class
- Implement opencl kernel for LHS non-transposed and RHS non-transposed
- Implement opencl kernel for LHS non-transposed and RHS transposed
- Add test fixture and dataset for matmul
- Implement transpose_tensor() for reference implementation to transpose high dimensional tensors
Resolves: COMPMID-5944, COMPMID-5951
Co-authored-by: Gunes Bayir <gunes.bayir@arm.com>
Co-authored-by: Ramy Elgammal <ramy.elgammal@arm.com>
Change-Id: I1d5b8978f41be27baddb3153ade880472141573f
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9333
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/gpu/cl/kernels/ClNativeMatMulKernel.h')
-rw-r--r-- | src/gpu/cl/kernels/ClNativeMatMulKernel.h | 70 |
1 files changed, 70 insertions, 0 deletions
diff --git a/src/gpu/cl/kernels/ClNativeMatMulKernel.h b/src/gpu/cl/kernels/ClNativeMatMulKernel.h new file mode 100644 index 0000000000..1cd74365df --- /dev/null +++ b/src/gpu/cl/kernels/ClNativeMatMulKernel.h @@ -0,0 +1,70 @@ +/* + * Copyright (c) 2023 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ACL_SRC_GPU_CL_KERNELS_CLNATIVEMATMULKERNEL +#define ACL_SRC_GPU_CL_KERNELS_CLNATIVEMATMULKERNEL + +#include "arm_compute/core/CL/CLHelpers.h" +#include "arm_compute/core/CL/CLKernelLibrary.h" +#include "arm_compute/core/KernelDescriptors.h" +#include "src/core/common/Macros.h" +#include "src/gpu/cl/ClCompileContext.h" +#include "src/gpu/cl/IClKernel.h" + +namespace arm_compute +{ +namespace opencl +{ +namespace kernels +{ +class ClNativeMatMulKernel : public IClKernel +{ +public: + ClNativeMatMulKernel(); + ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClNativeMatMulKernel); + /** Initialise the kernel's input and output. + * + * @param[in] compile_context The compile context to be used. + * @param[in] lhs Input tensor for the LHS matrix. Data type supported: F32/F16. + * Dimensions above 2 are collapsed onto dimension 2 and represent the batch. + * @param[in] rhs Input tensor for the RHS matrix. Data type supported: same as @p lhs. + * Dimensions above 2 are collapsed onto dimension 2 and represent the batch. + * @param[out] output Output tensor info. Data type supported: same as @p lhs + * @param[in] matmul_info Attributes for Batch MatMul Kernel + */ + void configure(const ClCompileContext &compile_context, ITensorInfo *lhs, ITensorInfo *rhs, ITensorInfo *output, const MatMulKernelInfo &matmul_info); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to @ref ClNativeMatMulKernel::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *lhs, const ITensorInfo *rhs, const ITensorInfo *output, const MatMulKernelInfo &matmul_info); + + // Inherited methods overridden: + void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; +}; +} // namespace kernels +} // namespace opencl +} // namespace arm_compute +#endif /* ACL_SRC_GPU_CL_KERNELS_CLNATIVEMATMULKERNEL */ |