From 4abc9d1a842e90162afe5349e3d51298fa0b8af4 Mon Sep 17 00:00:00 2001 From: SiCong Li Date: Wed, 28 Oct 2020 14:19:28 +0000 Subject: COMPMID-3730 Remove padding from CLGEMMMatrixMultiplyKernel Patch1 * Remove default definition for STORE_BLOCK_BOUNDARY_AWARE to avoid elusive bugs * Clean up gemm_mm_interleaved* and gemm_mm_floating_point* kernels * Relocate to gemm_v1.cl to avoid clashing with new kernels * Rename compile time arguments to conform with the established terminology(MNKB), and to facilitate the use of STORE_BLOCK_BOUNDARY_AWARE Change-Id: Ia85c746b2536cad87257a79685b459b5d2f9a1be Signed-off-by: SiCong Li Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4329 Tested-by: Arm Jenkins Reviewed-by: Gian Marco Iodice Comments-Addressed: Arm Jenkins --- src/core/CL/CLKernelLibrary.cpp | 24 ++++++++++++++---------- 1 file changed, 14 insertions(+), 10 deletions(-) (limited to 'src/core/CL/CLKernelLibrary.cpp') diff --git a/src/core/CL/CLKernelLibrary.cpp b/src/core/CL/CLKernelLibrary.cpp index 8c5607eebc..33f0da0aa1 100644 --- a/src/core/CL/CLKernelLibrary.cpp +++ b/src/core/CL/CLKernelLibrary.cpp @@ -206,16 +206,16 @@ const std::map CLKernelLibrary::_kernel_program_map = { "gemm_ma_f32", "gemm.cl" }, { "gemm_mv", "gemv.cl" }, { "gemm_mv_quantized", "gemv.cl" }, - { "gemm_mm_interleaved_transposed_f16", "gemm.cl" }, - { "gemm_mm_interleaved_transposed_f16_acc32", "gemm.cl" }, - { "gemm_mm_interleaved_transposed_f16_bifrost", "gemm.cl" }, - { "gemm_mm_interleaved_transposed_f32", "gemm.cl" }, - { "gemm_mm_interleaved_transposed_f32_bifrost", "gemm.cl" }, - { "gemm_mm_floating_point", "gemm.cl" }, - { "gemm_mm_floating_point_f16_bifrost", "gemm.cl" }, - { "gemm_mm_floating_point_f16_bifrost_acc32", "gemm.cl" }, - { "gemm_mm_floating_point_f32_bifrost", "gemm.cl" }, - { "gemm_mm_floating_point_f32_bifrost_1000", "gemm.cl" }, + { "gemm_mm_interleaved_transposed_f16", "gemm_v1.cl" }, + { "gemm_mm_interleaved_transposed_f16_acc32", "gemm_v1.cl" }, + { "gemm_mm_interleaved_transposed_f16_bifrost", "gemm_v1.cl" }, + { "gemm_mm_interleaved_transposed_f32", "gemm_v1.cl" }, + { "gemm_mm_interleaved_transposed_f32_bifrost", "gemm_v1.cl" }, + { "gemm_mm_floating_point", "gemm_v1.cl" }, + { "gemm_mm_floating_point_f16_bifrost", "gemm_v1.cl" }, + { "gemm_mm_floating_point_f16_bifrost_acc32", "gemm_v1.cl" }, + { "gemm_mm_floating_point_f32_bifrost", "gemm_v1.cl" }, + { "gemm_mm_floating_point_f32_bifrost_1000", "gemm_v1.cl" }, { "gemm_mm_native", "gemm.cl" }, { "gemm_mm_reshaped_lhs_nt_rhs_t", "gemm.cl" }, { "gemm_mm_reshaped_lhs_nt_rhs_t_texture", "gemm.cl" }, @@ -688,6 +688,10 @@ const std::map CLKernelLibrary::_program_source_map = { "gemm.cl", #include "./cl_kernels/gemm.clembed" + }, + { + "gemm_v1.cl", +#include "./cl_kernels/gemm_v1.clembed" }, { "gemmlowp.cl", -- cgit v1.2.1