From 538a076aae14dbc1940c52057e135f8a1872aa11 Mon Sep 17 00:00:00 2001 From: Giorgio Arena Date: Fri, 14 May 2021 16:38:43 +0100 Subject: Fix oclgrind error on CLGEMMLowp reshaped only RHS quantized per channel Fix corner case in which the quantization is per channel and OFM == 1 The function can safely set the per_channel quantization flag to false since there is only one output channel This way, the kernel will avoid adding useless padding to the output multipliers and shifts Resolve COMPMID-4384 Change-Id: Ic03452bfaf52d1be536cd371721adedd2e580a08 Signed-off-by: Giorgio Arena Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5648 Tested-by: Arm Jenkins Reviewed-by: Georgios Pinitas --- .../CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp | 2 +- src/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.cpp | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp b/src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp index d39900a561..37c11000db 100644 --- a/src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp +++ b/src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp @@ -256,7 +256,7 @@ std::pair validate_and_configure_window(ITensorInfo *input0, ITe window_changed = window_changed || update_window_and_padding(win_out, bias_access); } - if(output_multipliers != nullptr && output_multipliers->dimension(0) > 1) + if(output_multipliers != nullptr && output_stage.is_quantized_per_channel) { AccessWindowHorizontal output_multipliers_access(output_multipliers, 0, num_elems_processed_per_iteration_x); AccessWindowHorizontal output_shifts_access(output_shifts, 0, num_elems_processed_per_iteration_x); diff --git a/src/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.cpp b/src/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.cpp index b189955c04..5a9ff7990f 100644 --- a/src/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.cpp +++ b/src/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.cpp @@ -344,6 +344,12 @@ void CLGEMMLowpMatrixMultiplyCore::configure(const CLCompileContext &compile_con GEMMLowpOutputStageInfo gemmlowp_output_stage = gemm_info.gemmlowp_output_stage(); gemmlowp_output_stage.output_data_type = _matrix_a->info()->data_type(); + if(num_filters == 1) + { + // Per-channel quantization with OFM == 1 is equivalent to uniform quantization. + // Setting this flag to false prevents the kernel from adding useless padding to the output multipliers and shifts + gemmlowp_output_stage.is_quantized_per_channel = false; + } gemm_kernel_info.output_stage = gemmlowp_output_stage; -- cgit v1.2.1