diff options
| author | Gunes Bayir <gunes.bayir@arm.com> | 2022-08-10 14:31:23 +0100 |
|---|---|---|
| committer | Ramy Elgammal <ramy.elgammal@arm.com> | 2022-08-11 09:37:32 +0100 |
| commit | 9937c83908e4c1b783912927e45f613ecce0c879 (patch) | |
| tree | 28dac173ca56b39982821765a6801e75cd1fbc35 | |
| parent | 1524a0275dbe29005b7518dbe2991834b7e908d7 (diff) | |
| download | ComputeLibrary-9937c83908e4c1b783912927e45f613ecce0c879.tar.gz | |
Fix CTS/SLTS failure related to Depthwise Convolution
The issue is caused by GPUTarget not being set explicitly
for the Depthwise convolution kernel, but it's being used
in its build configuration. This causes the default value
to be used and enables some unsafe FP optimizations.
Resolves: COMPMID-5490
Change-Id: I5300a1168962cacb62cf49db795f052cf6740c7e
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8059
Reviewed-by: SiCong Li <sicong.li@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
| -rw-r--r-- | src/runtime/CL/functions/CLDepthwiseConvolutionLayer.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/runtime/CL/functions/CLDepthwiseConvolutionLayer.cpp b/src/runtime/CL/functions/CLDepthwiseConvolutionLayer.cpp index 2ee23c4262..e821726d0e 100644 --- a/src/runtime/CL/functions/CLDepthwiseConvolutionLayer.cpp +++ b/src/runtime/CL/functions/CLDepthwiseConvolutionLayer.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2021 Arm Limited. + * Copyright (c) 2017-2022 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -227,6 +227,7 @@ void CLDepthwiseConvolutionLayer::configure(const CLCompileContext &compile_cont const ConvolutionInfo conv_kernel_info{ conv_info, depth_multiplier, act_info, dilation }; + _dwc_native_kernel->set_target(gpu_target); _dwc_native_kernel->configure(compile_context, input_to_use, weights_to_use, biases, output_to_use, dwc_native_compute_info, conv_kernel_info, output_multipliers_to_use, output_shifts_to_use); |