diff options
| author | Milos Puzovic <milos.puzovic@arm.com> | 2024-03-28 13:28:21 +0000 |
|---|---|---|
| committer | Milos Puzovic <milos.puzovic@arm.com> | 2024-04-04 11:15:37 +0000 |
| commit | 1e91d71aaadd2154a708de1bc95f3c937b1e718a (patch) | |
| tree | 0c7da8e5d5c8dab609587b73602187496fa734d7 | |
| parent | 77bbe2e08b0376edfd3f504950be7f4b5720eeb0 (diff) | |
| download | ComputeLibrary-1e91d71aaadd2154a708de1bc95f3c937b1e718a.tar.gz | |
Parallelise im2col along dimensions with higher number of iterations
Signed-off-by: Milos Puzovic <milos.puzovic@arm.com>
Change-Id: I362f3f4a42e218424fca917bed22003ec9d5609c
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11363
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
| -rw-r--r-- | src/cpu/operators/CpuGemmConv2d.cpp | 13 |
1 files changed, 10 insertions, 3 deletions
diff --git a/src/cpu/operators/CpuGemmConv2d.cpp b/src/cpu/operators/CpuGemmConv2d.cpp index 7460f2020c..55d950ff4a 100644 --- a/src/cpu/operators/CpuGemmConv2d.cpp +++ b/src/cpu/operators/CpuGemmConv2d.cpp @@ -809,9 +809,16 @@ void CpuGemmConv2d::run(ITensorPack &tensors) if (!_skip_im2col) { // Run input reshaping - unsigned int y_dim = get_data_layout_dimension_index(_data_layout, DataLayoutDimension::HEIGHT); - ITensorPack pack = {{TensorType::ACL_SRC, src}, {TensorType::ACL_DST, im2col_output.get()}}; - NEScheduler::get().schedule_op(_im2col_kernel.get(), y_dim, _im2col_kernel->window(), pack); + unsigned int hint_dim = get_data_layout_dimension_index(_data_layout, DataLayoutDimension::HEIGHT); + unsigned int x_dim = get_data_layout_dimension_index(_data_layout, DataLayoutDimension::WIDTH); + unsigned int hint_dim_iterations = _im2col_kernel->window().num_iterations(hint_dim); + unsigned int x_dim_iterations = _im2col_kernel->window().num_iterations(x_dim); + if (hint_dim_iterations < NEScheduler::get().num_threads() && x_dim_iterations > hint_dim_iterations) + { + hint_dim = x_dim; + } + ITensorPack pack = {{TensorType::ACL_SRC, src}, {TensorType::ACL_DST, im2col_output.get()}}; + NEScheduler::get().schedule_op(_im2col_kernel.get(), hint_dim, _im2col_kernel->window(), pack); gemm_input_to_use = im2col_output.get(); } |