aboutsummaryrefslogtreecommitdiff
path: root/src/core/CL
AgeCommit message (Collapse)Author
2018-11-02COMPMID-477 - Optimized batched case in CLConvolutionLayerGian Marco Iodice
Change-Id: I4ef18f49f1da0cb816aaa0762466b940792c15ed Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84162 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-417: Port PoolingLayer to new validation.Georgios Pinitas
Change-Id: I7f2f5f5f81ad9932661fc4c660bf90614288bc96 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/85270 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-513 Choose maximum local workgroup size at run timesteniu01
Change-Id: I9ab3cf6dc92a93b0ae5f746e078355e443b3a545 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84906 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-452 CL Depthwise Separable Convolution Layer kernel implementation, ↵Giorgio Arena
validation and benchmarking for 3x3xC depthwise filter and DataType::F32. Change-Id: I95c0c87709763cdbf58d0de66025eac86e30791b Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82768 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Steven Niu <steven.niu@arm.com>
2018-11-02COMPMID-477 - Optimized CLNormalizationLayerGian Marco Iodice
CLPixelWiseMultiplication has been removed within the function Change-Id: Ibe7edd7921d5cef6ff68fdeeca89771129a8eaea Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84459 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2018-11-02COMPMID-431 Port OpenCL pooling layer to use fixed pointsteniu01
Change-Id: I6a73cd6582097aaefa83588aad789bdefdc74406 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79967 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
2018-11-02COMPMID-477 - Optimizing Pooling 3x3 with stride_x <= 3 on OpenCLGian Marco Iodice
Change-Id: Ie000166307cdb5bfae00ebf84d35e49a6bfb9dbd Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83372 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Pablo Tello <pablo.tello@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-417: Cleanup CL FullyConnectedLayerMoritz Pflanzer
Change-Id: Ic7191be1f136c6aad4037cf2ec4bc6d7d0e440d3 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83713 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-417 - Fixed call of direct convolution 1x1 for bifrostGian Marco Iodice
Change-Id: Ic4e56e8881b8c66758e67c486514ec397cf43f8e Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84592 Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-11-02COMPMID-477 - Optimized Direct Convolution 3x3 and 5x5 (f32) for Bifrost.Gian Marco Iodice
Each work-item computes 4x3 output elements in case of 3x3 convolution and 4x2 in case of 5x5 convolution Change-Id: I6ebbaff8b7e971c1f90d5845c0b58d2a40f39df5 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84345 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-11-02COMPMID-417: Add in-place support for batch-normalization.Georgios Pinitas
Change-Id: I4b0c9348f3bc2addc198a76fadd1b583abf42b60 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84434 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
2018-11-02COMPMID-450 Add YOLOV2 benchmark testsSiCong Li
* Migrate BatchNormalizationLayer to new benchmark system. * Add YOLOV2 benchmark tests. * Fix F16 type issue in activation_layer cl kernel. * Separate precommit tests from nightly tests. Change-Id: I3f206e3f7469be6749d630ede8dcc9fb399de8b0 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81582 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-478 Implemnt CL direct convolution 5x5steniu01
Change-Id: I4b975aff310cda9964d8c5dcee182d5d5c82741b Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83474 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
2018-11-02COMPMID-474 - Add support for QS8/QS16 DirectConvolution CLMichalis Spyrou
Change-Id: I537e4acbc02c8d880ff8630ea62223e0f1a1dda3 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82875 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Pablo Tello <pablo.tello@arm.com>
2018-11-02COMPMID-424 NEON/CL Harris Corners validation tests.Giorgio Arena
Change-Id: I82d2a73f515a8d45d16b9ddb702fea51ae05c82e Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79687 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
2018-11-02COMPMID-417 - Fixed bug in CLCol2ImKernek related to the stride passed ↵Gian Marco Iodice
during the configuration Change-Id: I9818f72e5ddd0d21f6700c651fc968ff61507424 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83909 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
2018-11-02COMPMID-417 - Added clFinish to CLSymbolsGian Marco Iodice
Change-Id: If3ee89d91f105489c766b9e714fdf72da8fbfe78 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83664 Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-11-02COMPMID-459 Collapse CL Im2col's higher dimensionssteniu01
Change-Id: I0ccc39cbcf6926e6810faf3fe264c4af7adc3f7b Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83070 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-11-02COMPMID-477 - Optimizing CLDirectConvolution 3x3 on OpenCL and added the ↵Gian Marco Iodice
auto configuration Change-Id: I3c8384dcbc9d7786943134bb658dafb35356d90d Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83253 Reviewed-by: Steven Niu <steven.niu@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-415: Add autoconfigure to CLCol2ImKernelAnthony Barbier
Change-Id: I50c114d0c78d443a21bf43aa36a370474f0769ce Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82955 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
2018-09-17COMPMID-417: Fix CLNormalization error issue.Georgios Pinitas
Change-Id: Ie538245ee0451e4cdb28120e80b9a65f56a07e7d Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82933 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-472 : Implement Floor for CL and NEON.Georgios Pinitas
Change-Id: I675a4545b1fe9ab665a07c834720bfe7ff589cee Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82527 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-443 collapse higher dimension for CL col2im kernelsteniu01
Change-Id: I99d41c7c95b8d4e3cd5c1685c68936b6a2db4192 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81885 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417 NEON/CL MeanStdDev bugfix using FillBorderKernelGiorgio Arena
Change-Id: Ic48ba7f69783d0e1e80611264e2bc67d1732436e Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81293 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-355 Implement CL DirectConvolution1x1SiCong Li
* Add FP16 to validation tests. * Complete benchmark tests for CL and NEON Direct Convolution. Change-Id: Ie73d8580832372db01b82b39786fd9c8be560090 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82014 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-438: Add support for floating point Min-Max Location layer.Michele Di Giorgio
Change-Id: I84ae564a40fc7320a6f94a84d53906ba51404f51 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79797 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-413: Add support for QS8 and QS16 CLNormalizationLayer.Michele Di Giorgio
Change-Id: I1aaa9fb8d05796bbca9cfae584e084646552bb71 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80155 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417 - Bug Fix WarpPerspective kernelIsabella Gottardi
Change-Id: Ic26fb3b1b60c1a1f4848d683862a25bd1ebc2cc8 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82053 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Steven Niu <steven.niu@arm.com>
2018-09-17COMPMID-457 Fix F16 NormalizationLayer CL kernelSiCong Li
Change-Id: I307c7ef6a49c852615c4425dc8dc0b1066a6974f Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81895 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417: Fix CL F16 ActivationLayerMoritz Pflanzer
Change-Id: I5b703d59661f6b50e3de4b2313be381b8d27030a Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81951 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-455 - Optimizing CLIm2ColKernelGian Marco Iodice
Change-Id: Iee618948cc8f310ee9af2d786240e8120e4c6ab9 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81665 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-417: Fix F16 CLSoftmaxLayerMoritz Pflanzer
Change-Id: I231b1fcaea8bfb11f8306bc71fdde78fadeed60d Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81832 Reviewed-by: Steven Niu <steven.niu@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-415: Use half_float library for F16Moritz Pflanzer
3RDPARTY_UPDATE Change-Id: Iee572e18d5b1df71300d738cc8690f49d7203d5c Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81353 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-355 Implement 3x3 CL direct convolutionsteniu01
Change-Id: I1b44dc375045964e65557f0ead57a7c12d6bf097 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81418 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-417: Allow loading of custom OpenCL libraryMoritz Pflanzer
Change-Id: Ic26266f164ec26f9f2923c0ffe366fea37b970ae Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80090 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-446: Add support for QS8/QS16 CL Arithmetic Add/SubMichele Di Giorgio
Change-Id: I84fc457a9c28856a11322944822d2fabaf92e8e4 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80528 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
2018-09-17COMPMID-425 Port CLBatchnormalization to support QS8/QS16Michalis Spyrou
Change-Id: I46c93305f377666ea0915ff789b7dfdfff596087 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/78862 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417: Add Leaky RELU support for both NEON/CL.Georgios Pinitas
-Adds parametrizable leaky relu (x>0) ? x : a*x. Change-Id: Ief19a435b5832a30b56f4aaaf55125787addee94 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80575 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-443 Collapse higher dimension for pooling layer and normalization layersteniu01
Change-Id: Icd08eefbd938c11c77dc4264af1fa3664fb336bc Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80568 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-443 Change CLSoftMaxLayerKernel to use 3D tensor and collapse the ↵steniu01
higer dimension Change-Id: I730ef45d855113d8baa7d89818441e168ea43c63 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80573 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-406: Port CLActivationLayer to use QS8/QS16.Georgios Pinitas
Change-Id: Ia4114984c38e1d2027ad97335b3c6c11f5754e23 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/78727 Reviewed-by: Anthony Barbier <anthony.barbier@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417: Port DepthConcatenate to QS8/QS16 for NEON/CL.Georgios Pinitas
Change-Id: I3dddae63043c7aa18d908a4fc8abacf3c64f98ca Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80081 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Steven Niu <steven.niu@arm.com>
2018-09-17COMPMID-443 Use 3D tensor for pixel multiply (Needed for Normalization Layer)Anthony Barbier
Change-Id: I117688f12334e6afc705c863acdf71b0bb1fc6e8 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80352 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2018-09-17COMPMID-443: Collapse higher dimensions for activation layerAnthony Barbier
Change-Id: I5943235aff1bb6440e3ab08e818d53aa5d94143a Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80349 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2018-09-17COMPMID-443: Use 3D tensors for fill_border_imageAnthony Barbier
2x performance improvement on some GoogLeNet Pooling tests Change-Id: If75336aa6308731a06462a73cd9209d24574509e Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80342 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Steven Niu <steven.niu@arm.com> Reviewed-by: Pablo Tello <pablo.tello@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
2018-09-17COMPMID-431 Port CLDepthConvert to use 8-bit and 16-bit fixed pointsteniu01
Change-Id: Iedea9e985427e6242f34a5362615f79c0526d5bd Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79786 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-417 Checking CL non uniform support at runtime.steniu01
What have been done in the ticket are: 1. Add support to check whether cl-non-unform-workgroup is supported at runtime 2. Add helper function to check the CL version at runtime 3. Add boolen to check whether CLSecheduler's init has been called. Change-Id: I6e6df8eb5cebfac7229aa406242bb183477fd191 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80265 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
2018-09-17COMPMID-429: Port CLSoftmaxLayer to QS16.Georgios Pinitas
Change-Id: I3a0394364629654747439372d32f692b6ca29ee0 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/80219 Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
2018-09-17COMPMID-440, COMPMID-441 - Port CLConvolutionLayer and CLFullyConnectedLayer ↵Gian Marco Iodice
to support 16 bit fixed point Change-Id: I8d8ef2cb5ec453eb83fba8d8077550b96ed4bceb Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79837 Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
2018-09-17COMPMID-409: Add support for QS8 and QS16 CLPixelWiseMultiplication.Michele Di Giorgio
Change-Id: I7f66d49d746ba9fb6e726ccab83d3a97b8ddef80 Reviewed-on: http://mpd-gerrit.cambridge.arm.com/78491 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
generated by cgit v1.2.1 (git 2.23.0) at 2024-07-06 12:29:44 +0000