From d9c1d44185e4fea72ce4d37d76021237fe64eeaf Mon Sep 17 00:00:00 2001 From: Adnan AlSinan Date: Fri, 6 Oct 2023 19:07:48 +0100 Subject: Port MatMul to Dynamic Fusion + CKW boilerplate code - Port Matmaul to to Dynamic Fusion. - Prepare a CKW boilerplate code. - Implement the following classes: - MatMulAttributes - GPUMatMulSettings - GpuMatMul - ClComponentMatMul - GpuCkwMatMul Signed-off-by: Adnan AlSinan Change-Id: I5a7c183b293973e8a4233b554b2affe0bb28f44d Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10453 Comments-Addressed: Arm Jenkins Reviewed-by: Jakub Sujak Reviewed-by: Viet-Hoa Do Tested-by: Arm Jenkins Benchmark: Arm Jenkins --- .../sketch/attributes/MatMulAttributes.h | 62 ++++++++++ .../sketch/gpu/operators/GpuMatMul.h | 132 +++++++++++++++++++++ 2 files changed, 194 insertions(+) create mode 100644 arm_compute/dynamic_fusion/sketch/attributes/MatMulAttributes.h create mode 100644 arm_compute/dynamic_fusion/sketch/gpu/operators/GpuMatMul.h (limited to 'arm_compute') diff --git a/arm_compute/dynamic_fusion/sketch/attributes/MatMulAttributes.h b/arm_compute/dynamic_fusion/sketch/attributes/MatMulAttributes.h new file mode 100644 index 0000000000..fc512e9ff9 --- /dev/null +++ b/arm_compute/dynamic_fusion/sketch/attributes/MatMulAttributes.h @@ -0,0 +1,62 @@ +/* + * 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_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_ATTRIBUTES_MATMULATTRIBUTES_H +#define ACL_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_ATTRIBUTES_MATMULATTRIBUTES_H + +#include "arm_compute/core/Types.h" + +namespace arm_compute +{ +namespace experimental +{ +namespace dynamic_fusion +{ +/** Attributes are backend-agnostic parameters (in addition to the input/output tensors) of an operator. + */ + +/** MatMul attributes */ +class MatMulAttributes +{ +public: + /* Get adjoint LHS flag value (transpose LHS before multiplication) */ + bool adj_lhs() const; + + /* Get adjoint RHS flag value (transpose RHS before multiplication) */ + bool adj_rhs() const; + + /* Set adjoint LHS flag value (transpose LHS before multiplication) */ + MatMulAttributes adj_lhs(bool adj_lhs); + + /* Set adjoint RHS flag value (transpose RHS before multiplication) */ + MatMulAttributes adj_rhs(bool adj_rhs); + +private: + bool _adj_lhs{false}; + bool _adj_rhs{false}; +}; +} // namespace dynamic_fusion +} // namespace experimental +} // namespace arm_compute + +#endif // ACL_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_ATTRIBUTES_MATMULATTRIBUTES_H diff --git a/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuMatMul.h b/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuMatMul.h new file mode 100644 index 0000000000..6ba5b1db93 --- /dev/null +++ b/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuMatMul.h @@ -0,0 +1,132 @@ +/* + * 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_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUMATMUL_H +#define ACL_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUMATMUL_H + +#include "arm_compute/core/Error.h" +#include "arm_compute/core/KernelDescriptors.h" +#include "arm_compute/dynamic_fusion/sketch/attributes/MatMulAttributes.h" + +namespace arm_compute +{ +namespace experimental +{ +namespace dynamic_fusion +{ +/** Forward declaration */ +class GpuWorkloadContext; +class GpuWorkloadSketch; + +/** Operator backend specific settings +*/ +class GpuMatMulSettings +{ +public: + /** Set N0: number of columns processed by each work-item */ + GpuMatMulSettings &n0(int n0); + /** Get N0: number of columns processed by each work-item */ + int n0() const; + + /** Set M0: number of rows processed by each work-item */ + GpuMatMulSettings &m0(int m0); + /** Get M0: number of rows processed by each work-item */ + int m0() const; + + /** Set K0: number of inner accumulations */ + GpuMatMulSettings &k0(int k0); + /** Get K0: number of inner accumulations */ + int k0() const; + +private: + int _n0{0}; /**< Number of columns processed by each work-item */ + int _m0{0}; /**< Number of rows processed by each work-item */ + int _k0{0}; /**< Number of inner accumulations */ +}; + +/** Operator interface. */ +class GpuMatMul final +{ +public: + /* Attributes are a set of backend-agnostic parameters that define what an operator does */ + using Attributes = MatMulAttributes; + /* Settings are a set of backend-specific parameters that influence the implementation of a operator */ + using Settings = GpuMatMulSettings; + + /* Create an operator and fuse it into the workload sketch. + * @note If @ref validate_op() fails, the creation also fails and may throw an error. + * @note If @ref validate_op() fails, @p sketch remains unchanged and valid. + * + * Valid data type configurations: + * |LHS |RHS |dst | + * |:--------------|:--------------|:--------------| + * |F16 |F16 |F16 | + * |F32 |F32 |F32 | + * + * + * @param[in,out] sketch Workload sketch into which the operator will be fused + * @param[in] lhs Input tensor info 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 info for the RHS matrix. Data type supported: same as @p lhs. Dimensions above 2 are collapsed onto dimension 2 and represent the batch. + * @param[in] attributes Operator attributes + * @param[in] settings Operator settings + * + * @return Pointer for the destination tensor info + */ + static ITensorInfo *create_op(GpuWorkloadSketch &sketch, + ITensorInfo *lhs, + ITensorInfo *rhs, + const Attributes &attributes, + const Settings &settings); + + /* Check if the operator configuration is supported, irrespective of fusion + * + * @param[in] context Workload context within which the operator is running + * @param[in] lhs Input tensor info for the LHS matrix. + * @param[in] rhs Input tensor info for the RHS matrix. + * @param[in] attributes Operator attributes + * @param[in] settings Operator settings + * + * @return Status + */ + static Status is_supported_op(const GpuWorkloadContext &context, + const ITensorInfo *lhs, + const ITensorInfo *rhs, + const Attributes &attributes, + const Settings &settings); + + /* Check if the operator configuration is supported and if it can be fused into the workload sketch. + * + * Parameters are similar to @ref GpuMatMul::create_op() + * + * @return Status + */ + static Status validate_op(const GpuWorkloadSketch &sketch, + const ITensorInfo *lhs, + const ITensorInfo *rhs, + const Attributes &attributes, + const Settings &settings); +}; +} // namespace dynamic_fusion +} // namespace experimental +} // namespace arm_compute +#endif // ACL_ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUMATMUL_H -- cgit v1.2.1