aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2022-04-01Add CPU Pool3d FP16/32 implementationmasterAdnan AlSinan
- Add implementation for the CPU pooling 3d layer. - NDHWC data layout support - Support FP32/FP16. - Add Pool3d to the operator list. - Fix CL Pool3d kernel comments to generate the operator list. Resolves: COMPMID-4671 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: I92478a154beb12541525b648ed3dd5a58c8f27fa Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7311 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-29NEQLSTM: Add support for QASYMM8_SIGNED for input_to_forget_weightsPablo Marquez Tello
* QLSTM only supports QSYMM8 for the argument input_to_forget_weights * We add support for QASYMM8_SIGNED by dequantizing and requantizing to QSYMM8 * Resolves COMPMID-5184 Change-Id: I1cae18d81dafdb7ae722b520a1354cf4a56b9606 Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7321 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-28Enable -static-openmp for newer NDK supportGiorgio Arena
Android NDKs >= r21 need the '-static-openmp' flag enabled when using openmp. The check on the NDK version is performed by querying the compiler binary's clang version. NDK versions <= r20 use clang 8 and below. Versions >= r21 use clang 9 and above Resolves: COMPMID-5052 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I910f8d7a0a59866c06b16c1a856ea877bcbd7ad6 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7344 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-24QLSTM add support for different qinfo in weightsPablo Marquez Tello
Resolves COMPMID-5185 Change-Id: I61e1453e8851ab84c1cadc10587ebd23fd94799e Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7330 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-24[Dynamic Fusion] Implement build options generationGiorgio Arena
Resolves: COMPMID-5153 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: Ic34cc1f0d092fafa7c2faa4dd705cf8f68eaf87e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7317 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-03-24Add cleanup comments in SConscriptGiorgio Arena
Resolves: COMPMID-5151 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: Ia5d7e4910bfe219cc13241615abc54ac886016b9 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7270 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-03-17Setup automatic kernel window for dynamic fusionGunes Bayir
The window of the fused kernels is equal to the root kernel's window. For this reason, the following kernels does not change the window and intermediate tensors have the same shape and info as the destination tensor. Resolves: COMPMID-5152 Change-Id: I25fe2fab8304ecaabfc2e4ade9bbf31a600a5033 Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7316 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com>
2022-03-16Remove deprecated interface from arm_compute.Francesco.Petrogalli@arm.com
The function `get_gemm_method` in arm_compute is deprecated in favor of the method `arm_gemm::GemmCommon<TypeInput, TypeOutput>::get_config()`. Signed-off-by: francesco.petrogalli@arm.com Change-Id: Idd5d879180c3995d5a07a727aa9216b8f94f01ba Signed-off-by: Francesco.Petrogalli@arm.com <francesco.petrogalli@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7304 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com>
2022-03-16Fix dynamic fusion test Scons to enable unit testsSiCong Li
Partially resolves: COMPMID-5152 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: I3b584e33d160284132e62ecfc8ea80dbd7b08a1d Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7306 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-03-15Implementation of ClPooling3dramelg01
- For NDHWC layout - For F16 and F32 data types - Mixed Precision stil not supported Resolves: COMPMID-4670 Signed-off-by: ramy.elgammal@arm.com Change-Id: I0e14a13e4625569e8e5ee67e6033bd1efe0da469 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7262 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-03-10Added windows native build supportPablo Tello
Resolves MLCE-739 Signed-off-by: Pablo Tello <pablo.tello@arm.com> Change-Id: I30a11393e928061c82a5c93d8ec195c04a0e838b Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7279 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@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>
2022-03-08Decouple fuseBatchNormalizationKernelYair Schwarzbaum
- Decouple data type for CPU implementation supported data types are: fp32, fp16 Resolves COMPMID-4613 Signed-off-by: Yair Schwarzbaum <yair.schwarzbaum@arm.com> Change-Id: I8aff3ba2d446f64e4d182a866e3a3debc9ef613b Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7175 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-08Merge kernel prototype patchGiorgio Arena
Resolves: COMPMID-5151 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: Ic4024d5cd4819fe917a1d49621f1866ae2e90a37 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7260 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-03Removing SVE / SVE2 guards from decoupled kernelsalerah01
Jira: COMPMID-5172 Signed-off-by: alerah01 <alex.rahlis@arm.com> Change-Id: I1b9ace8e573f85830f29728a27adfe39a0cab113 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7241 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-03Decouple castKernelYair Schwarzbaum
Resolves: COMPMID-4625 Signed-off-by: Yair Schwarzbaum <yair.schwarzbaum@arm.com> Change-Id: I3c30f007804b179e5e2b439f421fbd4e57fb02e1 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7149 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
2022-03-01Add Pool3d reference implementationGunes Bayir
This patch - adds the reference implementation for the 3D pooling layer - supports FP32/FP16 and INT8/UINT8 types - adds a function to calculate the output shape for 3D pooling - adds a new type for describing pool 3d info (Pool3DInfo) Resolves: COMPMID-4659 Change-Id: I22a18fa30625c98fa827ef1b50781db6893ba9c4 Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7219 Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-03-01Multi ISA Technical DebtDana Zlotnik
* Update json struct meet multi-ISA updates * Add impl.cpp in kernels where we only have impl.h Resolves COMPMID-5173 Change-Id: I5da3c4b016a5d0115c4ba46cbfefde7bce518ac1 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7191 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-01Unify aarch64 arch names in documentationGunes Bayir
This patch addresses the confusion in our documentation and errata by unifying different aliases for our architecture names. For example, armv8a and arm64-v8a refers to the same arch, as armv8.2-a and arm64-v8.2-a are. Resolves: COMPMID-4997 Change-Id: Id08f7a27363abbd2de75b2db96a91417e559417a Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7224 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Pablo Marquez Tello <pablo.tello@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>
2022-02-22Add disclaimer of ACL duty to protect the memory of the process using it ↵ramelg01
from hacking Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com> Resolves: COMPMID-2132 Change-Id: I76945425244202edb4aad03c1269e69c9f2608ae Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7036 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
2022-02-22Decouple CpuDirectConv2dKernelalerah01
Resolves COMPMID-4626 Exclude SVE & SVE2 paths from android.bp NDK version does not support these extensions. Change-Id: I49b147d2a84819975d3225f2920106fa1a0d742f Signed-off-by: alerah01 <alex.rahlis@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7136 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
2022-02-21Adding Comments to ComputeLibrary.dir on expected file structureMohammed Suhail Munshi
Resolves: [COMPMID-4976] Change-Id: I9f2d08f706a3fa68613b8c6937fcbe8aba70983b Signed-off-by: Mohammed Suhail Munshi <MohammedSuhail.Munshi@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7160 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
2022-02-21Decouple CpuDepthwiseConv2dNativeKernelDana Zlotnik
Resolves COMPMID-4632 Change-Id: I5e2a9f0f7801a2afaa35de871ab29cd7238923fd Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7115 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-21Fix performance regression on Arm(R) Mali(TM)-G71Gian Marco Iodice
Resolves COMPMID-5095 Change-Id: Ib7f6e891095372f870fc08833cca922909506dad Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7158 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-17Decouple NEL2NormalizeLayerKernelYair Schwarzbaum
Resolves: COMPMID-4615 Signed-off-by: Yair Schwarzbaum <yair.schwarzbaum@arm.com> Change-Id: Iadbfb3e45831a5072962b5b9f61e8ae2e674ccc4 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7016 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-16Update Errata with ticket numbersAdnan AlSinan
Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: Ia084965f90d4117cbc86baeb6e28703395aaa77d Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7154 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-16Fixed threshould argument order in NE/CL/LSTMPablo Marquez Tello
* Fixed hardcoded LOGISTIC activation in ACL reference * Partially resolves MLCE-60 * Resolves COMPMID-5139 Change-Id: I50e75339084ea53bf75acf18aa3e5cdafcf34c15 Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7150 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-16Fix FP16 for validation in multi_isa buildsGiorgio Arena
Resolves COMPMID-5133 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: Ice2ed60aa91042e1dc78a3548f8b960f185cb6e5 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7120 Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-15Update ReadmeAdnan AlSinan
Resolve COMPMID-5087 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: Id59bcee79fcce2d481f7dfe04b78fab1371dfdd2 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7128 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
2022-02-15Fix Doxygen buildAdnan AlSinan
- Remove NEMaxUnpoolingLayerKernel referance from the change log. Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: I77bd3ce31196499a9b48345da0632a99fa11740a Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7137 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-14Exclude SVE & SVE2 paths from android.bpDana Zlotnik
NDK version does not support these extensions. Resolves COMPMID-4921 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Change-Id: Icd506c00a5cecafafd682d07c32690ccf8ca587f Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/c/VisualCompute/ComputeLibrary/+/397247 Tested-by: bsgcomp <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7131 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2022-02-14Decouple CpuGemmMatrixMultiplyKernel and CpuGemmMatrixAdditionKernelDana Zlotnik
Resolves COMPMID-4629, COMPMID-4631 Change-Id: Idceafc84735116ef63ec13a202895f954b87e32f Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7095 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-14Port MaxUnpoolingLayer kernel and add KernelSelect vaidation testDana Zlotnik
Resolves COMPMID-4958 Change-Id: Ibed5155f2e3ece46635f6ea9617bf11cefc402b1 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7028 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-14Add kernel selection UT for submitted kernelsDana Zlotnik
* Softmax kernel * Elementwise unary kernel * Elementwise binary ** This change require some refactor in the kernel cpp and h files Resolves COMPMID-5043 Change-Id: I58979b023ec31d759690847b3f85fc4baefbbf98 Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7033 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
2022-02-11Update Release NotesAdnan AlSinan
- Update Release notes - Update Minimum toolchains requirements Resolves COMPMID-5088 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: I5d1a9e287e0407fa4c7ab5b53cf0f3082cbb251e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7099 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-11Inclusive language updateAdnan AlSinan
- Update Github README.md - Add section about inclusive language in our documention Resolves COMPMID-5129 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: Iab56d284a81061a1e62f4df7163553fdf47c2d7d Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7100 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-11Improve start-up time for concatenation layersramelg01
- pass tensor's dimensions at runtime rather than compile time - Add guard macro to compile only kernel of internest Resolves: COMPMID-5121 Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com> Change-Id: I76b7c0cf56d803f58ebff5494c904ace2a86ef5a Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7097 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-10Int32 to Float32 implicit casting fixMohammed Suhail Munshi
- Int32 is cast to Float32 implicitly during range checking comparisons, value may change due to float32 range limitations. - Fixed by casting float32 var to double64 before comparisons. - This issue prevented NDK r23b compilation Resolves: [COMPMID-5051] Change-Id: I1695f9ad5fa147d81cbcfda9e0152aefab960c82 Signed-off-by: Mohammed Suhail Munshi <MohammedSuhail.Munshi@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7093 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com>
2022-02-10Fix performance regression on the first layer of convolution-based modelGian Marco Iodice
Change-Id: Ia6d7282ee0c91c14dc71505a9f72f7d412d12055 Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7088 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
2022-02-10Improve start-up time for winograd_output_transform_*_nhwcramelg01
- pass tensor's dimensions at runtime rather than compile time - Add guard macro to compile only kernel of internest Resolves: COMPMID-5120 Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com> Change-Id: I87c3b56ce0cd3c97ffdeabdd9c5d433f361bb005 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7101 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-09Remove deprecated remap functions.Adnan AlSinan
- Remove CLRemapKernel. - Remove NERemapKernel. Partially resolves COMPMID-4984 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: Ia61f9ac7447695d81178701cf0e9b7625a91eccc Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7056 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-09Improve start-up time for winograd_input_transform_*_nhwcramelg01
- pass tensor's dimensions at runtime rather than compile time - Add guard macro to compile only kernel(s) of internest Resolves: COMPMID-5119 Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com> Change-Id: Ib01098e397011a1201c2800c62a8954ec70e63e8 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7083 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-08Improve start-up time for winograd_filter_transform_*_nhwcramelg01
- pass tensor's dimensions at runtime rather than compile time - Add guard macro to compile only kernel of internest Resolves: COMPMID-5118 Signed-off-by: Ramy Elgammal <ramy.elgammal@arm.com> Change-Id: Ie42c3c07fdd817ce62e7cad354381bc22c6e9264 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7058 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-08Fix build for gcc 9Giorgio Arena
Resolves COMPMID-5127 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I2b1eefdf5d9a58a13295b27f2d66e461d9b9289c Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7086 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
2022-02-08Update SONAME_VERSION in SConscriptAdnan AlSinan
Partially resolves COMPMID-5090 Signed-off-by: Adnan AlSinan <adnan.alsinan@arm.com> Change-Id: Ia53db44c3ae5e07fe01742d622b5331aa678acf6 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7057 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
2022-02-07Adding support for Arm Mali-G710Gian Marco Iodice
Resolves COMPUTE-13901 Change-Id: Ib83d737066a55ab6452bdc34e3e4cba2d466d72a Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6971 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-04SCons build system refactoring (phase #2).Motti Gondabi
* Add kernel selection at build time * Modify filelist.json to allow files separation as part of our kernel decoupling process. Issues to address after this change will be merged: (1) Remove SVE/SVE2 defines from already decoupled kernels (2) Adapt the new file list structure (filelist.json) resolves COMPMID-4996 and COMPMID-5048 Change-Id: I8c17a9d6b150bbc7d8c1f2ed38060be82b6aa904 Signed-off-by: Motti Gondabi <motti.gondabi@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7006 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: SiCong Li <sicong.li@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-03Added a multithreaded test for CL backend.Pablo Marquez Tello
* Resolves MLCE-662 Change-Id: Ic85ddcf6608d440ec0b143736a5f4a031a4ca237 Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6507 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-02-02Revert "Rework gemm_mm_reshaped_only_rhs_ kernels with new macros"Ramy Elgammal
This reverts commit 10e88a7351 "Rework gemm_mm_reshaped_only_rhs_ kernels with new macros" Resolves: COMPMID-5095 Signed-off-by: Ramy Elgammal<ramy.elgammal@arm.com> Change-Id: I46e167882f072e7508b6101d295accb6e089e740 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7045 Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>