aboutsummaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2021-02-23Avoid division with zero on CLGEMM lhs/rhs configurationGeorgios Pinitas
Resolves COMPMID-4285 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: I414e3ba352ed86345d6355d9ae366ea92e641842 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5153 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-22Comply with Trademark rules for use of Neon, Arm and MaliSheri Zhang
Full trademarks available in README.md Resolves: COMPMID-4257 Signed-off-by: Sheri Zhang <sheri.zhang@arm.com> Change-Id: Ibfba2adf2eef3449433f467464ebd87d7198474d Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5116 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-19Free _gemm_output_3d's memory region after run()Giorgio Arena
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I689a9b6ad5a4486316363674491c594e153b19fb
2021-02-18Set CLDirectConvolutionLayerKernel NCHW _border_size to input paddingGiorgio Arena
Change-Id: I5802c470683647b7426b3b6e7d17280cabc32163 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5100 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-18Set _gemm_output_3d's padding as the output's in NEGEMMConvolutionLayerGiorgio Arena
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I31a51e8d28262680f9bb6f7530e28385efcb6b3b Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5090 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-17Update deprecation noticesGiorgio Arena
Change-Id: I6db0a10249b2212add3ff4bb2598b03cae2bff55 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5098 Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-16Handle Conv2d layer with implicit output padding in NHWCGeorgios Pinitas
Corner cases exist when output top/bottom padding is non-zero for Convolution Layer. This can cause invalid output from the NEGEMMConvolutionLayer as assembly kernel integration does not efficiently handles such cases. As a workaround we always allocate a memory-managed auxiliary tensor which we use as an output for GEMM when padding exists and then we copy to the padded output. If no padding exists we import the output tensor memory to the temporary buffer and perform calculation as we did before. Resolves: COMPMID-4114 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: If82d0e115b8369b91d775895d5315b044306cc74 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5083 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-16Fix data layout retention and handling of leftovers when there is no paddingMichele Di Giorgio
Resolves COMPMID-4258 Change-Id: I8782bf725cd0d376d538021406eb1f5be962b2cb Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/c/VisualCompute/ComputeLibrary/+/298627 Reviewed-by: Teresa Charlin Reyes <teresa.charlinreyes@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Teresa Charlin Reyes <teresa.charlinreyes@arm.com> Tested-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5082 Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com>
2021-02-11Building android-nn-driver failedManuel Bottini
Properly include CLGEMMHelpers Resolves: COMPMID-4255 Change-Id: I06f0da8b7b51e5abdff1eb883a5dae4138a68f89 Signed-off-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5054 Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-11Validate mlgo gemm type selection and fall back to default heuristicsSiCong Li
GEMM kernel type returned by mlgo heuristics in each of the CLGEMM and CLGEMMLowpMatrixMultiplyCore could also be invalid. Fix this by falling back to default heuristics, similar to how we deal with gemm configs for now. Resolves COMPMID-3847 Change-Id: Iae7c1dcd7def04969ad13a4c132873fda8c8a571 Signed-off-by: SiCong Li <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5044 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
2021-02-10Revert changes on tensor's strides and fix CLDepthwiseConvolution 3x3 QuantizedGiorgio Arena
- Revert changes in strides > num_dimensions. Set them to 0 - Fix offset calculcation in depthwise 3x3 quantized using select and stride_y for max offset Resolve COMPMID-4254 Change-Id: Ia99b9637f18b99b1fa3d4b7b4892046027d3e7e5 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5040 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-10Integrate MLGO into CLGEMMLowpMatrixMultiplyCore for native kernelSiCong Li
Resolves COMPMID-3846 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: Iad66f6dd7fa5b13ebace9f95fbc2fc4d677cf6a9 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5032 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-10Decouple CpuPoolingKernel data type and data layoutSheri Zhang
1. Decouple data layout for CpuPoolingKernel: NCHW & NHWC 2. Decouple data type for CpuPoolingKernel NHWC Partially solves: COMPMID-3999 Signed-off-by: Sheri Zhang <sheri.zhang@arm.com> Change-Id: I3c6535eebdddeb467b7c68a7287a16959b5b9695 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5039 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Integrate MLGO into CLGEMM for reshaped kernelSiCong Li
Resolves COMPMID-3845 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: I878ea6dc076177095816a75f9bc951326fd095b3 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5031 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Make Softmax kernels and operator statelessMichalis Spyrou
COMPMID-3997 Change-Id: I3a3cc76d8247dd769d9a5e6e171d718ea909312c Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4986 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Add mlgo to graph examplesSiCong Li
Resolves COMPMID-3847 Change-Id: I99f73bfc8eda66e8ce1dd1f2a18be76e9d826569 Signed-off-by: SiCong Li <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5033 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Fix CLDepthwiseConvolutionLayer 3x3 QASYMM8Giorgio Arena
Fix errors when computing tensors with one element only - Replace Tensor3D with raw pointers so to get rid of offset to first element for NCHW layout - Add stronger out of bound constraints for NHWC layout - Set the border size to the input's padding for NHWC - Fill the strides == 0 with the largest stride, so to avoid accessing empty strides and multiplying by 0 Resolve COMPMID-4088 Change-Id: I751a4e6d7094b3c42306ff7f53af848fd35f19ac Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5024 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part4SiCong Li
Apply cl_gemm::auto_heuristics to CLGEMMLowpMatrixMultiplyCore for the selection of gemm config reshaped only rhs and gemm kernel type Resolves: COMPMID-3843, COMPMID-3844 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: I351c76b052a1e52acec23a217bb111da8e40518e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4992 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part3SiCong Li
* Implement a common interface across both functions and across mlgo and default heuristics. This interface is implemented as: * A set of adaptor functions in new cl_gemm::auto_heuristics namespace as: * select_default_*: For selecting configs using default heuristics * select_mlgo_*: For selecting configs using mlgo heuristics These adaptor functions have the same interface * On top of these adaptor functions, a set of auto_select_* functions that automatically selects between mlgo and default (prioritize mlgo). Note that auto_select_gemm_config_* are implemented in each individual function. This is because the auto selection depends on the validation of its hosting functions. When we are able to decouple and abstract the validation logics, it's possible to share the core auto_gemm_config_* in cl_gemm::auto_heuristics namespace as well. * Apply this interface in CLGEMM for the selection of gemm config reshaped only rhs and gemm kernel type Resolves: COMPMID-3843, COMPMID-3844 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: Idf7fb46837a027449aae1e251346b2701866309a Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4991 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part2SiCong Li
* Associate CLScheduler with CLGEMMHeuristicsHandle * Add option in arm_compute_validation for mlgo file path * Extend logging for the selection of gemm configurations Resolves: COMPMID-3843, COMPMID-3844 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: I869c3a4122414ae6a7bbd721966c1da37621ca11 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5002 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part1SiCong Li
* Create a new public handle class CLGEMMHeuristicsHandle It is responsible for the loading and lifetime management of the underlying heuristics * Add to_string utility to several mlgo constructs for logging Resolves: COMPMID-3843, COMPMID-3844 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: Ib9c65e076daa6a9a204999cde9abf366dbabc496 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5001 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-09ArmNN internal OOB failed to build ACL with the use of a deleted functionManuel Bottini
Remove "noexcept" from move assign constructor to support gcc 5.4.0 Resolves: COMPMID-4241 Change-Id: Idff6b8a7468e4733ed6cc87a910bd85590463de0 Signed-off-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5025 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-08Make memset/copy functions state-lessSheri Zhang
Port following functions: - CLCopy - CLFill - CLPermute - CLReshapeLayer - CLCropResize Resolves: COMPMID-4002 Signed-off-by: Sheri Zhang <sheri.zhang@arm.com> Change-Id: I8392aa515aaeb5b44dab6122be6a795d08376d5f Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5003 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-08Make NEON Pooling kernels and functions state-lessMichele Di Giorgio
Partially resolves COMPMID-3999 Change-Id: Ib39d40694df5c5f0a9401488e0c3af3ac26e8c55 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4984 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-04Tweak scheduling use of SQDMULH in quantized AVG poolingMichele Di Giorgio
Resolves COMPMID-4195 Change-Id: Ie5116c1ddddccafba40432fd4b5245bb27890a88 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4997 Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-04Fix CL Compiler frontend failure for ArgMinMaxGiorgio Arena
- Change select condition's data type to satisfy its signature - Add failing test case with VEC_SIZE == 1 Resolve: COMPMID-4110 Change-Id: I52287bff7a2108f92fd12164e267df6c074d5508 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4978 Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-03Fix OpenCL direct convolutionGian Marco Iodice
- The ARM DOT macro was using wrong variables for performing the dot product - K0 could be a non power of 2 values when IFM was not a multiple of 16 - Refactor the test for direct convolution NHWC Resolves COMPMID-4135, COMPMID-4155 Change-Id: I3a2dc89ef613ae20245cfc28e76ea36c55eaf81d Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4962 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-03Fix segfault in fsrcnn.tflite in GpuAccTeresa Charlin
* In CLDirectConvolution check for non-bias separately Resolves: COMPMID-4214 Change-Id: I83c0688e9b48d059665bbc6e1f0f050a516132d6 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4980 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-02-03Make CL Pooling kernels and functions state-lessMichele Di Giorgio
Resolves COMPMID-4000 Change-Id: I64878f93c033b4928fdefbb964c37c67fdecfaab Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4971 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2021-02-03Add WBSM tuning to CLTunerManuel Bottini
Add WBSM as possible parameter to be tuned Add helper functions to check WBSM support and setting the value in the kernel Update tuning parameter lists to use WBSM Update CLTuner to use WBSM The WBSM tuning is exposed as a parameter to be set at compile time by setting the CLTuningInfo CLTuningInfo contains information about the tuning mode and if wbsm tuning enabled Resolves: COMPMID-3936 Change-Id: Id53697c9c6d2cef41c049f368002f6197351b3ed Signed-off-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4914 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2021-02-02Implement MLGO moduleSiCong Li
* Implement MLGOHeuristics which provides a query and a loading interface * Implement a top-down parser MLGOParser for parsing dotmlgo * Add validation tests for MLGOHeuristics Resolves COMPMID-3840, COMPMID-3841 Signed-off-by: SiCong Li <sicong.li@arm.com> Change-Id: Iae96d2779524b2dd83623d1a3a30ef57823ae084 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4941 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-02Account for eventual padding in Pooling assembly kernelsMichele Di Giorgio
Resolves: COMPMID-4217 Change-Id: I8426a494ab0d3c6479de977cdab803d35bbe1e40 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4967 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-02-01Compress OpenCL kernel files using zlib for AndroidGeorgios Pinitas
Kernel files are embedded into the binary as the default option when building which leads to binary size bloating. Add `compress_kernels` option and utilize zlib for further compressing the text kernel files and reduce the overall binary size. We use a base64 encoding/decoding to ensure that the strings can be easily embedded. This adds to the binary size but still the overall reduction is significant. Maximum compression level 9 is used. Option is currently restricted to Android builds as android toolchain provides a zlib library. Initial experimentations indicate a binary size reduction of 50% Resolves: COMPMID-4017 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: Iee81b8c00391b26a5f41642699692928a4d6bd6e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4958 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
2021-02-01Make data_layout an attribute of the Scale functionMichele Di Giorgio
Resolves COMPMID-4208 Change-Id: I61ca670134a005462ad0528a5aff9507a90860e7 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4942 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-27Make CL Elementwise Unary kernels and functions state-lessMichele Di Giorgio
Resolves COMPMID-4004 Change-Id: I1dfe8bc52c1ff394ea208ba98b51033c738746a4 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4922 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2021-01-27Add broadcasting support in NEPixelWiseMultiplicationKernel for FP16Pablo Marquez Tello
* FP16 kernel missed the broadcast path * Resolves: COMPMID-4113 Change-Id: I8fd79030f2ae9c268dfeed672a57c6d0f64f58f4 Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4926 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-27Force early exit of threads when output_height=1 and channels=1 in assembly ↵Michele Di Giorgio
pooling Resolves COMPMID-4156, COMPMID-4125 Change-Id: I408e51e9759448e6190490975ccab2c1bc07a813 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4927 Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-27Make Gpu Elementwise Operations kernels and functions state-lessMichele Di Giorgio
Partially resolves COMPMID-4004 Change-Id: Ie99b6c676af5be4c2a67affc06fec378f3dff8ec Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4918 Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-27Fix NEON/CL CustomConvolution failuresManuel Bottini
Properly handle UNDEFINED border mode as done in the reference Change-Id: I23b117be3ee5cd0e566f6161681d38493627fc9b Resolves: COMPMID-4133 Signed-off-by: Manuel Bottini <manuel.bottini@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4919 Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-26Make CLArithmeticSubtraction kernel and function state-lessMichele Di Giorgio
Resolves COMPMID-4008 Change-Id: Ic5f40610e771f31e6d301dfae976c81e9c79fa8b Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4917 Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-26Remove iostream header file when not neededGian Marco Iodice
Change-Id: Ic6b3e94fc252c2d6fe17e6bc32900d11c68ae399 Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4916 Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-26Make CLArithmeticAddition kernel and function state-lessMichele Di Giorgio
Resolves COMPMID-4006 Change-Id: Iddc32b0b250142aac9a4a7b9dc0eef462d196025 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4913 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com>
2021-01-26Rename functions/classes for elementwise unary operationsSang-Hoon Park
* Create CpuElementwiseUnary operator * Rename kernel classes * Make the kernels stateless Partially implements: COMPMID-4003 Change-Id: Ie0440cd01d4924847d6991b4df7ccaf311439297 Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4912 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-25CTS Failures in Android CpuAcc in AddTeresa Charlin
* Fusing the activation into Addition is not supported in CpuAcc. Armnn uses this validate function to decide if fusion is supported. Resolves COMPMID-4121 Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com> Change-Id: I78cec326057eada32c2bf7d0e8d64a4af5854cd9 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4911 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-22Rename functions/classes for elementwise operationsSang-Hoon Park
* Create CpuElementwise operator * Rename kernel classes * Make the kernels stateless Partially implements: COMPMID-4003 Change-Id: I4ef9c61a3acc3ac5dbe46463d62dcb88a5face21 Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4881 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-22Fix Scale CTS failuresSheri Zhang
- Move back to use template for NEON implementaion to be more mantainable Resolves: COMPMID-4112 Signed-off-by: Sheri Zhang <sheri.zhang@arm.com> Change-Id: I27b60674d743e3d7ec5bb065898cb75cb40bcf0e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4900 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
2021-01-22Check for unsupported armv7a optimized pooling kernelsMichele Di Giorgio
Resolves COMPMID-4136 Change-Id: I3c1f9e615cb2d2c48464ac5b7dbf0a26dc435889 Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4906 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-22Fixed segfault NEDeconvLayer due to overflow.Pablo Marquez Tello
* The CTS test TestRandomGraph/RandomGraphTest#SmallGraph_TENSOR_FLOAT16_Rank4/43 segfaulted when running on CpuAcc + 8-2a. * Added a guard against overflow in DeconvLayer::validate * Resolves: COMPMID-4115 Change-Id: I6687b67dbde3921a2795224e8107a9add9d358ff Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4890 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2021-01-22Update Fp16 heuristic for Mali-G76 and Mali-G77Gian Marco Iodice
Change-Id: I199c10574d0d84737adbc73829f60a220c7088b4 Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4897 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-22CTS failures in Android R and Q in GpuAcc in ArgMinMaxGiorgio Arena
- Fix ambiguosity with select in OpenCL - Define a new macro for signed integer data type of the same input data type's size. This is needed because some ops (e.g. logical operators) in OpenCL work in this way Resolves: COMPMID-4116, COMPMID-4110 Signed-off-by: Giorgio Arena <giorgio.arena@arm.com> Change-Id: I560eda63fce24abd03d061f78f2f2ca951053fd0 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4898 Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
generated by cgit v1.2.1 (git 2.23.0) at 2024-05-19 18:47:08 +0000