aboutsummaryrefslogtreecommitdiff
path: root/src/cpu/kernels/internal/CpuPool2dAssemblyWrapperKernel.cpp
AgeCommit message (Collapse)Author
2022-04-25Update Neon™ pooling kernelramelg01
- Reduce duplication and simplify overall structure. - Improve multi-threaded performance by sharing more data in lower-level caches. Partially Resolves: COMPMID-5054 Signed-off-by: Ramy Elgammal<ramy.elgammal@arm.com> Change-Id: I5f4dc50913401d5c1cbfc10b866fae9490cbc4d7 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7404 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Andrew Mundy Reviewed-by: Sheri Zhang <sheri.zhang@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-09Revert mws heuristics for CpuPool2dGiorgio Arena
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I194a78244124b72ace9a9e4630001b94c46ac74e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7185 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-22Fix CpuPool2d regression on A53/A55 due to mwsGiorgio Arena
Resolves COMPMID-5045 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I85562a96cc6998567a96b6f43d1a0145f7c7b330 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7177 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-12-25Add tests for FP Cpu Pooling where pool region is completely outside the inputSiCongLi
* Add floating point validation tests for this configuration * Fix reference implementation to return -inf for this configuration * Prohibit this config in Cl, as well as non-float cases in Cpu * Direct this config to non-asm path Resolves COMPMID-4998 Change-Id: If88025c51b14ea337aea2441c548f858e95e5819 Signed-off-by: SiCongLi <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6857 Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-12-17Redirect Pool2d unsupported cases to unoptimized pathGiorgio Arena
Unsupported case: Any of the convolution padding > pool size Resolve COMPMID-4999 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I4279f3f14d865e0faf53e04847b52ef328970cf2 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6817 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
2021-11-16Implement 1D Adaptive Workload Splitting in CPPSchedulerDana Zlotnik
Resolves COMPMID-4649 Change-Id: I941d2f8a40737ff05c49f6695a42884731ef2dc9 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6656 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-10-18Implement Minimum Workload Size (MWS) in all CPPKernels used by small networksDana Zlotnik
* create get_mws method in ICPPKernel class that retuns default value for all kernels * overwrite the default value for all the kernels used by small networks (according to banchmark case) Resolves COMPMID-4648 Change-Id: I46d7cae61217213279d2ee740edc73f600b6d576 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6412 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-08-25Move CPU/GPU files from Core/Runtime to the respective backend foldersGeorgios Pinitas
Legacy structure contained two libraries core/runtime with two backends in each. We reduce the core/runtime libraries to a single library thus merging the backend files Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: I69545765fe7a730368105cdbd067d3135ec7a174 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6155 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>