aboutsummaryrefslogtreecommitdiff
path: root/arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h
diff options
context:
space:
mode:
authorSiCongLi <sicong.li@arm.com>2020-03-02 15:39:15 +0000
committerSiCong Li <sicong.li@arm.com>2020-03-25 11:23:19 +0000
commit2e5fd637205770ec5e11096e6e19b8efc67d544e (patch)
treec7f5b53d6ec54cf39fcc638bad9d054c3e04bf43 /arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h
parent51a9558d484e812707a776efceef18b93698cec9 (diff)
downloadComputeLibrary-2e5fd637205770ec5e11096e6e19b8efc67d544e.tar.gz
COMPMID-3098 Fuse Relu and Bounded Relu with FullyConnected NEON
Change-Id: Id28062445590d6c06b35f7d7434eb38393ae94a7 Signed-off-by: SiCongLi <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2875 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h')
-rw-r--r--arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h14
1 files changed, 9 insertions, 5 deletions
diff --git a/arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h b/arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h
index 508159eb77..74dedcf4c5 100644
--- a/arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h
+++ b/arm_compute/runtime/NEON/functions/NEGEMMLowpMatrixMultiplyCore.h
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2017-2019 ARM Limited.
+ * Copyright (c) 2017-2020 ARM Limited.
*
* SPDX-License-Identifier: MIT
*
@@ -28,9 +28,12 @@
#include "arm_compute/core/NEON/INEKernel.h"
#include "arm_compute/core/NEON/kernels/NEConvertQuantizedSignednessKernel.h"
#include "arm_compute/core/NEON/kernels/NEConvertQuantizedSignednessKernel.h"
+#include "arm_compute/core/NEON/kernels/NEGEMMInterleave4x4Kernel.h"
+#include "arm_compute/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMLowpOffsetContributionKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMLowpOffsetContributionOutputStageKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMLowpReductionKernel.h"
+#include "arm_compute/core/NEON/kernels/NEGEMMTranspose1xWKernel.h"
#include "arm_compute/runtime/IFunction.h"
#include "arm_compute/runtime/IMemoryManager.h"
#include "arm_compute/runtime/MemoryGroup.h"
@@ -60,7 +63,7 @@ class NEGEMMLowpMatrixMultiplyCore : public IFunction
{
public:
/** Constructor */
- NEGEMMLowpMatrixMultiplyCore(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
+ NEGEMMLowpMatrixMultiplyCore(std::shared_ptr<IMemoryManager> memory_manager = nullptr, IWeightsManager *weights_manager = nullptr);
/** Prevent instances of this class from being copied (As this class contains pointers) */
NEGEMMLowpMatrixMultiplyCore(const NEGEMMLowpMatrixMultiplyCore &) = delete;
/** Default move constructor */
@@ -109,10 +112,11 @@ public:
private:
MemoryGroup _memory_group;
+ IWeightsManager *_weights_manager;
NEGEMMAssemblyDispatch _asm_glue;
- std::unique_ptr<INEKernel> _mm_kernel;
- std::unique_ptr<INEKernel> _mtx_a_reshape_kernel;
- std::unique_ptr<INEKernel> _mtx_b_reshape_kernel;
+ NEGEMMLowpMatrixMultiplyKernel _mm_kernel;
+ NEGEMMInterleave4x4Kernel _mtx_a_reshape_kernel;
+ NEGEMMTranspose1xWKernel _mtx_b_reshape_kernel;
NEGEMMLowpMatrixAReductionKernel _mtx_a_reduction_kernel;
NEGEMMLowpMatrixBReductionKernel _mtx_b_reduction_kernel;
NEGEMMLowpOffsetContributionKernel _offset_contribution_kernel;