diff options
author | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-26 17:23:59 +0100 |
---|---|---|
committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-27 14:06:27 +0000 |
commit | 4718706b1141d5cccb006a7f86d65c1fde6c54ff (patch) | |
tree | c35b4c3568d10981327abae9820accfab1ceabad /src/runtime/gpu/cl/operators/ClGemmConvolution.h | |
parent | 387f80ed5b4c8ef32cfdba4ea27734dd257f87f0 (diff) | |
download | ComputeLibrary-4718706b1141d5cccb006a7f86d65c1fde6c54ff.tar.gz |
Dispatch Conv2d using the Direct method when necessary
Direct Convolution will be preferred for large kernel sizes (>=5) and
when input feature maps are greater or equal to output feature maps.
Resolves: COMPMID-4708
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ic9fe8b178a58bd95e7c47597221e47d344f96c79
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5996
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/runtime/gpu/cl/operators/ClGemmConvolution.h')
0 files changed, 0 insertions, 0 deletions