Port CLGEMM to memory injecting interface
Moves the following kernels:
- CLGEMMMatrixMultiplyKernel
- CLGEMMMatrixMultiplyNativeKernel
- CLGEMMMatrixMultipluReshapedKernel
- CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
Moves the following functions
- CLGEMM
Introduces facilities to easy handling of auxiliary temporary buffers
under then new run interface. Such are:
- CLAuxTensorHandler: That allows wrapping of workspace buffers memory
to CLBuffer objects
- Ability to inject TensorInfo to allocator without transferring
ownership. This reduce the copy overhead if needed.
Resolves: COMPMID-4188
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I7055435d831b05b749b26302082e4ac45f26dfb0
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5498
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
diff --git a/docs/user_guide/release_version_and_change_log.dox b/docs/user_guide/release_version_and_change_log.dox
index 0f4d4cc..a975e8b 100644
--- a/docs/user_guide/release_version_and_change_log.dox
+++ b/docs/user_guide/release_version_and_change_log.dox
@@ -280,7 +280,7 @@
- CLLogits1DMaxShiftExpSumKernel
- CLLogits1DNormKernel
- CLHeightConcatenateLayerKernel
- - @ref CLGEMMMatrixMultiplyKernel
+ - CLGEMMMatrixMultiplyKernel
- @ref CLGEMMLowpQuantizeDownInt32ScaleKernel
- @ref CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel
- @ref CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel
@@ -567,14 +567,14 @@
The default "axis" value for @ref NESoftmaxLayer, @ref NELogSoftmaxLayer is changed from 1 to 0.
Only axis 0 is supported.
- The support for quantized data types has been removed from @ref CLLogSoftmaxLayer due to implementation complexity.
- - Removed padding requirement for the input (e.g. LHS of GEMM) and output in @ref CLGEMMMatrixMultiplyNativeKernel, @ref CLGEMMMatrixMultiplyReshapedKernel, @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel and @ref CLIm2ColKernel (NHWC only)
+ - Removed padding requirement for the input (e.g. LHS of GEMM) and output in CLGEMMMatrixMultiplyNativeKernel, CLGEMMMatrixMultiplyReshapedKernel, CLGEMMMatrixMultiplyReshapedOnlyRHSKernel and @ref CLIm2ColKernel (NHWC only)
- This change allows to use @ref CLGEMMConvolutionLayer without extra padding for the input and output.
- Only the weights/bias of @ref CLGEMMConvolutionLayer could require padding for the computation.
- - Only on Arm® Mali™ Midgard GPUs, @ref CLGEMMConvolutionLayer could require padding since @ref CLGEMMMatrixMultiplyKernel is called and currently requires padding.
- - Added support for exporting the OpenCL buffer object to the OpenCL image object in @ref CLGEMMMatrixMultiplyReshapedKernel and @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.
+ - Only on Arm® Mali™ Midgard GPUs, @ref CLGEMMConvolutionLayer could require padding since CLGEMMMatrixMultiplyKernel is called and currently requires padding.
+ - Added support for exporting the OpenCL buffer object to the OpenCL image object in CLGEMMMatrixMultiplyReshapedKernel and CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.
- This support allows to export the OpenCL buffer used for the reshaped RHS matrix to the OpenCL image object.
- - The padding requirement for the OpenCL image object is considered into the @ref CLGEMMReshapeRHSMatrixKernel.
- - The reshaped RHS matrix stores the weights when GEMM is used to accelerate @ref CLGEMMConvolutionLayer.
+ - The padding requirement for the OpenCL image object is considered into the CLGEMMReshapeRHSMatrixKernel.
+ - The reshaped RHS matrix stores the weights when GEMM is used to accelerate CLGEMMConvolutionLayer.
v20.05 Public major release
- Various bug fixes.
@@ -739,7 +739,7 @@
- Added QASYMM16 support for:
- @ref CLBoundingBoxTransform
- Added FP16 support for:
- - @ref CLGEMMMatrixMultiplyReshapedKernel
+ - CLGEMMMatrixMultiplyReshapedKernel
- Added new data type QASYMM8_PER_CHANNEL support for:
- CLDequantizationLayer
- @ref NEDequantizationLayer
@@ -749,7 +749,7 @@
- @ref CLDepthwiseConvolutionLayer
- @ref NEDepthwiseConvolutionLayer
- Added FP16 mixed-precision support for:
- - @ref CLGEMMMatrixMultiplyReshapedKernel
+ - CLGEMMMatrixMultiplyReshapedKernel
- CLPoolingLayerKernel
- Added FP32 and FP16 ELU activation for:
- @ref CLActivationLayer
@@ -813,9 +813,9 @@
- @ref CLSinLayer
- CLBatchConcatenateLayerKernel
- @ref CLDepthToSpaceLayerKernel / @ref CLDepthToSpaceLayer
- - @ref CLGEMMLowpMatrixMultiplyNativeKernel
+ - CLGEMMLowpMatrixMultiplyNativeKernel
- CLGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel
- - @ref CLGEMMMatrixMultiplyNativeKernel
+ - CLGEMMMatrixMultiplyNativeKernel
- CLMeanStdDevNormalizationKernel /CLMeanStdDevNormalizationLayer
- @ref CLSpaceToDepthLayerKernel / @ref CLSpaceToDepthLayer
- New examples:
@@ -862,7 +862,7 @@
- @ref CLFFTRadixStageKernel
- @ref CLFFTScaleKernel
- @ref CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel
- - @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
+ - CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
- CLHeightConcatenateLayerKernel
- @ref CLDirectDeconvolutionLayer
- @ref CLFFT1D
@@ -947,9 +947,9 @@
- @ref CLRsqrtLayer
- @ref CLExpLayer
- CLElementWiseUnaryLayerKernel
- - @ref CLGEMMReshapeLHSMatrixKernel
- - @ref CLGEMMReshapeRHSMatrixKernel
- - @ref CLGEMMMatrixMultiplyReshapedKernel
+ - CLGEMMReshapeLHSMatrixKernel
+ - CLGEMMReshapeRHSMatrixKernel
+ - CLGEMMMatrixMultiplyReshapedKernel
- @ref CLRangeKernel / @ref CLRange
- @ref CLUnstack
- @ref CLGatherKernel / @ref CLGather
@@ -1369,7 +1369,7 @@
v17.03 Sources preview
- New OpenCL kernels / functions:
- CLGradientKernel, CLEdgeNonMaxSuppressionKernel, CLEdgeTraceKernel / CLCannyEdge
- - GEMM refactoring + FP16 support: CLGEMMInterleave4x4Kernel, CLGEMMTranspose1xWKernel, @ref CLGEMMMatrixMultiplyKernel, CLGEMMMatrixAdditionKernel / @ref CLGEMM
+ - GEMM refactoring + FP16 support: CLGEMMInterleave4x4Kernel, CLGEMMTranspose1xWKernel, CLGEMMMatrixMultiplyKernel, CLGEMMMatrixAdditionKernel / @ref CLGEMM
- CLGEMMMatrixAccumulateBiasesKernel / @ref CLFullyConnectedLayer
- CLTransposeKernel / @ref CLTranspose
- CLLKTrackerInitKernel, CLLKTrackerStage0Kernel, CLLKTrackerStage1Kernel, CLLKTrackerFinalizeKernel / CLOpticalFlow