From 951d520f492348ce07085f701078adcb48a4c1a2 Mon Sep 17 00:00:00 2001 From: Giorgio Arena Date: Wed, 8 Sep 2021 13:26:06 +0100 Subject: Remove padding from ClGemmMatrixMultiplyReshapedOnlyRhsKernel Resolve COMPMID-4450 Signed-off-by: Giorgio Arena Change-Id: I6f280d5d66ec43fb5cb06c83fe15a1f227ad165d Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6232 Tested-by: Arm Jenkins Reviewed-by: Gian Marco Iodice Comments-Addressed: Arm Jenkins --- src/core/CL/cl_kernels/common/gemm.cl | 40 +++++++-------- .../ClGemmMatrixMultiplyReshapedOnlyRhsKernel.cpp | 60 +++++----------------- .../ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h | 2 +- 3 files changed, 35 insertions(+), 67 deletions(-) (limited to 'src') diff --git a/src/core/CL/cl_kernels/common/gemm.cl b/src/core/CL/cl_kernels/common/gemm.cl index 76e6c21ab6..87921f51fd 100644 --- a/src/core/CL/cl_kernels/common/gemm.cl +++ b/src/core/CL/cl_kernels/common/gemm.cl @@ -1096,6 +1096,9 @@ __kernel void gemm_mm_reshaped_only_rhs_t(IMAGE_DECLARATION(lhs), uint y = get_global_id(1); uint z = get_global_id(2); + const bool cond_y = y == 0; + const bool cond_x = ((x + 1) * N0 >= N); + #if defined(DUMMY_WORK_ITEMS) if((x * N0 >= N) || (y * M0 >= M)) { @@ -1250,7 +1253,7 @@ __kernel void gemm_mm_reshaped_only_rhs_t(IMAGE_DECLARATION(lhs), #if defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (get_global_id(0) * (uint)N0 * sizeof(DATA_TYPE)); - LOAD_BLOCK(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, 1, PARTIAL_STORE_N0, false, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(1, DATA_TYPE, bias, BETA); @@ -1262,7 +1265,7 @@ __kernel void gemm_mm_reshaped_only_rhs_t(IMAGE_DECLARATION(lhs), #else // defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (x * (uint)N0 * sizeof(DATA_TYPE)) + (COMPUTE_M0_START_ROW(y, M0, PARTIAL_STORE_M0) * bias_stride_y) + z * bias_stride_z; - LOAD_BLOCK(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(M0, DATA_TYPE, bias, BETA); @@ -1278,9 +1281,6 @@ __kernel void gemm_mm_reshaped_only_rhs_t(IMAGE_DECLARATION(lhs), ACTIVATION_BLOCK(M0, ACTIVATION_TYPE, DATA_TYPE, VEC_SIZE, c, A_VAL, B_VAL); #endif // defined(ACTIVATION_TYPE) - const bool cond_y = y == 0; - const bool cond_x = ((x + 1) * N0 >= N); - // Store output block STORE_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, c, dst_addr, dst_stride_y, zout, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); @@ -1392,6 +1392,9 @@ __kernel void gemm_mm_reshaped_only_rhs_t_texture(IMAGE_DECLARATION(lhs), uint y = get_global_id(1); uint z = get_global_id(2); + const bool cond_y = y == 0; + const bool cond_x = ((x + 1) * N0 >= N); + #if defined(DUMMY_WORK_ITEMS) if((x * N0 >= N) || (y * M0 >= M)) { @@ -1596,7 +1599,7 @@ __kernel void gemm_mm_reshaped_only_rhs_t_texture(IMAGE_DECLARATION(lhs), #if defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (get_global_id(0) * (uint)N0 * sizeof(DATA_TYPE)); - LOAD_BLOCK(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, 1, PARTIAL_STORE_N0, false, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(1, DATA_TYPE, bias, BETA); @@ -1608,7 +1611,7 @@ __kernel void gemm_mm_reshaped_only_rhs_t_texture(IMAGE_DECLARATION(lhs), #else // defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (x * (uint)N0 * sizeof(DATA_TYPE)) + (COMPUTE_M0_START_ROW(y, M0, PARTIAL_STORE_M0) * bias_stride_y) + z * bias_stride_z; - LOAD_BLOCK(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(M0, DATA_TYPE, bias, BETA); @@ -1624,9 +1627,6 @@ __kernel void gemm_mm_reshaped_only_rhs_t_texture(IMAGE_DECLARATION(lhs), ACTIVATION_BLOCK(M0, ACTIVATION_TYPE, DATA_TYPE, VEC_SIZE, c, A_VAL, B_VAL); #endif // defined(ACTIVATION_TYPE) - const bool cond_y = y == 0; - const bool cond_x = ((x + 1) * N0 >= N); - // Store output block STORE_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, c, dst_addr, dst_stride_y, zout, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); @@ -1813,6 +1813,9 @@ __kernel void gemm_mm_reshaped_only_rhs_nt(IMAGE_DECLARATION(lhs), uint y = get_global_id(1); uint z = get_global_id(2); + const bool cond_y = y == 0; + const bool cond_x = ((x + 1) * N0 >= N); + #if defined(DUMMY_WORK_ITEMS) if((x * N0 >= N) || (y * M0 >= M)) { @@ -1992,7 +1995,7 @@ __kernel void gemm_mm_reshaped_only_rhs_nt(IMAGE_DECLARATION(lhs), #if defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (get_global_id(0) * (uint)N0 * sizeof(DATA_TYPE)); - LOAD_BLOCK(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, 1, PARTIAL_STORE_N0, false, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(1, DATA_TYPE, bias, BETA); @@ -2004,7 +2007,7 @@ __kernel void gemm_mm_reshaped_only_rhs_nt(IMAGE_DECLARATION(lhs), #else // defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (x * (uint)N0 * sizeof(DATA_TYPE)) + (COMPUTE_M0_START_ROW(y, M0, PARTIAL_STORE_M0) * bias_stride_y) + z * bias_stride_z; - LOAD_BLOCK(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(M0, DATA_TYPE, bias, BETA); @@ -2020,9 +2023,6 @@ __kernel void gemm_mm_reshaped_only_rhs_nt(IMAGE_DECLARATION(lhs), ACTIVATION_BLOCK(M0, ACTIVATION_TYPE, DATA_TYPE, VEC_SIZE, c, A_VAL, B_VAL); #endif // defined(ACTIVATION_TYPE) - const bool cond_y = y == 0; - const bool cond_x = ((x + 1) * N0 >= N); - // Store output block STORE_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, c, dst_addr, dst_stride_y, zout, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); @@ -2130,6 +2130,9 @@ __kernel void gemm_mm_reshaped_only_rhs_nt_texture(IMAGE_DECLARATION(lhs), uint y = get_global_id(1); uint z = get_global_id(2); + const bool cond_y = y == 0; + const bool cond_x = ((x + 1) * N0 >= N); + #if defined(DUMMY_WORK_ITEMS) if((x * N0 >= N) || (y * M0 >= M)) { @@ -2301,7 +2304,7 @@ __kernel void gemm_mm_reshaped_only_rhs_nt_texture(IMAGE_DECLARATION(lhs), #if defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (get_global_id(0) * (uint)N0 * sizeof(DATA_TYPE)); - LOAD_BLOCK(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(1, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, 1, PARTIAL_STORE_N0, false, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(1, DATA_TYPE, bias, BETA); @@ -2313,7 +2316,7 @@ __kernel void gemm_mm_reshaped_only_rhs_nt_texture(IMAGE_DECLARATION(lhs), #else // defined(BROADCAST_BIAS) __global uchar *bias_addr = bias_ptr + bias_offset_first_element_in_bytes + (x * (uint)N0 * sizeof(DATA_TYPE)) + (COMPUTE_M0_START_ROW(y, M0, PARTIAL_STORE_M0) * bias_stride_y) + z * bias_stride_z; - LOAD_BLOCK(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero); + LOAD_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, bias, bias_addr, 0, bias_stride_y, zero, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); #ifndef UNIT_BETA SCALE_BLOCK(M0, DATA_TYPE, bias, BETA); @@ -2329,9 +2332,6 @@ __kernel void gemm_mm_reshaped_only_rhs_nt_texture(IMAGE_DECLARATION(lhs), ACTIVATION_BLOCK(M0, ACTIVATION_TYPE, DATA_TYPE, VEC_SIZE, c, A_VAL, B_VAL); #endif // defined(ACTIVATION_TYPE) - const bool cond_y = y == 0; - const bool cond_x = ((x + 1) * N0 >= N); - // Store output block STORE_BLOCK_BOUNDARY_AWARE(M0, N0, DATA_TYPE, c, dst_addr, dst_stride_y, zout, PARTIAL_STORE_M0, PARTIAL_STORE_N0, cond_y, cond_x); diff --git a/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.cpp b/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.cpp index 149c92b7a9..04c1cd66c9 100644 --- a/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.cpp +++ b/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.cpp @@ -24,11 +24,7 @@ #include "src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h" #include "arm_compute/core/CL/ICLTensor.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/TensorInfo.h" -#include "arm_compute/core/Utils.h" #include "arm_compute/core/utils/misc/ShapeCalculator.h" -#include "src/core/AccessWindowStatic.h" #include "src/core/CL/CLUtils.h" #include "src/core/CL/CLValidate.h" #include "src/core/helpers/AutoConfiguration.h" @@ -118,18 +114,15 @@ Status validate_arguments(const ITensorInfo *src0, const ITensorInfo *src1, cons return Status{}; } -std::pair validate_and_configure_window(ITensorInfo *src0, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, const GEMMLHSMatrixInfo &lhs_info, - const GEMMRHSMatrixInfo &rhs_info, const GEMMKernelInfo &gemm_info, ElementsProcessed &num_elements_processed) +Window validate_and_configure_window(ITensorInfo *src0, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, const GEMMLHSMatrixInfo &lhs_info, + const GEMMRHSMatrixInfo &rhs_info, const GEMMKernelInfo &gemm_info, ElementsProcessed &num_elements_processed) { + ARM_COMPUTE_UNUSED(src0, src1, src2); unsigned int &num_elems_processed_per_iteration_x = num_elements_processed[0]; unsigned int &num_elems_processed_per_iteration_y = num_elements_processed[1]; bool reinterpret_input_as_3d = gemm_info.reinterpret_input_as_3d; bool reinterpret_output_as_3d = gemm_info.depth_output_gemm3d != 0; - Window win{}; - Window win_out{}; - bool window_changed = false; - // In case both input and dst have to be reinterpreted as 3D tensors, // force reinterpret_input_as_3d and reinterpret_output_as_3d to be false. // This approach should only be used when the input/dst tensors have pad on the y direction @@ -138,9 +131,6 @@ std::pair validate_and_configure_window(ITensorInfo *src0, ITens reinterpret_output_as_3d = false; } - // dst tensor auto initialization if not yet initialized - auto_init_if_empty(*dst, src0->clone()->set_tensor_shape(misc::shape_calculator::compute_mm_shape(*src0, *src1, gemm_info))); - TensorInfo tmp_info(*dst); if(reinterpret_output_as_3d) @@ -156,28 +146,14 @@ std::pair validate_and_configure_window(ITensorInfo *src0, ITens num_elems_processed_per_iteration_x = rhs_info.n0; num_elems_processed_per_iteration_y = lhs_info.m0; - win = calculate_max_window(tmp_info, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); - win_out = calculate_max_window(*dst, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); - - if(src2 != nullptr) - { - const int bias_processed_per_iteration_x = num_elems_processed_per_iteration_x; - - AccessWindowStatic src2_access(src2, 0, 0, - ceil_to_multiple(src2->dimension(0), bias_processed_per_iteration_x), - src2->dimension(1)); - - window_changed = update_window_and_padding(win, src2_access); - } + Window win = calculate_max_window(tmp_info, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); // Collapse along the Z direction // This collapse needs to be here in order to tune the Z dimension of LWS - Window collapsed = win; const unsigned int dimension_to_collapse = std::min(static_cast(dst->num_dimensions()), 2u); - collapsed = win.collapse(win, dimension_to_collapse); + Window collapsed = win.collapse(win, dimension_to_collapse); - Status err = (window_changed) ? ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Insufficient Padding!") : Status{}; - return std::make_pair(err, collapsed); + return collapsed; } } // namespace @@ -187,7 +163,7 @@ ClGemmMatrixMultiplyReshapedOnlyRhsKernel::ClGemmMatrixMultiplyReshapedOnlyRhsKe } void ClGemmMatrixMultiplyReshapedOnlyRhsKernel::configure(const CLCompileContext &compile_context, - ITensorInfo *src0, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, float alpha, float beta, + const ITensorInfo *src0, const ITensorInfo *src1, const ITensorInfo *src2, ITensorInfo *dst, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMKernelInfo &gemm_info) { ARM_COMPUTE_ERROR_ON_NULLPTR(src0, src1, dst); @@ -201,7 +177,10 @@ void ClGemmMatrixMultiplyReshapedOnlyRhsKernel::configure(const CLCompileContext _export_to_cl_image = rhs_info.export_to_cl_image; _has_pad_y = gemm_info.has_pad_y; - auto padding_info = get_padding_info({ src0, src1, dst }); + // dst tensor auto initialization if not yet initialized + auto_init_if_empty(*dst, src0->clone()->set_tensor_shape(misc::shape_calculator::compute_mm_shape(*src0, *src1, gemm_info))); + + auto padding_info = get_padding_info({ src0, src1, src2, dst }); // In case both input and dst have to be reinterpreted as 3D tensors, // force reinterpret_input_as_3d and reinterpret_output_as_3d to be false. @@ -218,9 +197,9 @@ void ClGemmMatrixMultiplyReshapedOnlyRhsKernel::configure(const CLCompileContext ElementsProcessed num_elements_processed{}; // Configure kernel window - auto win_config = validate_and_configure_window(src0, src1, src2, dst, lhs_info, rhs_info, gemm_info, num_elements_processed); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); - ICLKernel::configure_internal(win_config.second); + Window win = validate_and_configure_window(src0->clone().get(), src1->clone().get(), (src2 != nullptr) ? src2->clone().get() : nullptr, dst->clone().get(), lhs_info, rhs_info, gemm_info, + num_elements_processed); + ICLKernel::configure_internal(win); // If _reinterpret_input_as_3d = reinterpret_output_as_3d = true, // we will dispatch a batched-GEMM to reduce the complexity of the address calculation within the OpenCL kernel. @@ -314,18 +293,7 @@ Status ClGemmMatrixMultiplyReshapedOnlyRhsKernel::validate(const ITensorInfo *sr const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMKernelInfo &gemm_info) { - ElementsProcessed num_elements_processed{}; ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src0, src1, src2, dst, alpha, beta, lhs_info, rhs_info, gemm_info)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(src0->clone().get(), - src1->clone().get(), - src2 != nullptr ? src2->clone().get() : nullptr, - dst->clone().get(), - lhs_info, - rhs_info, - gemm_info, - num_elements_processed) - .first); - return Status{}; } diff --git a/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h b/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h index 3be96d3add..cb82b4af5e 100644 --- a/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h +++ b/src/gpu/cl/kernels/ClGemmMatrixMultiplyReshapedOnlyRhsKernel.h @@ -75,7 +75,7 @@ public: * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices */ void configure(const ClCompileContext &compile_context, - ITensorInfo *src0, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, float alpha, float beta, + const ITensorInfo *src0, const ITensorInfo *src1, const ITensorInfo *src2, ITensorInfo *dst, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMKernelInfo &gemm_info); /** Static function to check if given info will lead to a valid configuration * -- cgit v1.2.1