xref: /aosp_15_r20/external/armnn/src/backends/neon/workloads/NeonPreluWorkload.cpp (revision 89c4ff92f2867872bb9e2354d150bf0c8c502810)
1 //
2 // Copyright © 2017 Arm Ltd and Contributors. All rights reserved.
3 // SPDX-License-Identifier: MIT
4 //
5 
6 #include "NeonPreluWorkload.hpp"
7 #include "NeonWorkloadUtils.hpp"
8 
9 #include <aclCommon/ArmComputeUtils.hpp>
10 #include <armnn/utility/PolymorphicDowncast.hpp>
11 
12 #include <arm_compute/runtime/NEON/functions/NEPReluLayer.h>
13 
14 namespace armnn
15 {
16 
NeonPreluWorkloadValidate(const TensorInfo & input,const TensorInfo & alpha,const TensorInfo & output)17 arm_compute::Status NeonPreluWorkloadValidate(const TensorInfo& input,
18                                               const TensorInfo& alpha,
19                                               const TensorInfo& output)
20 {
21     const arm_compute::TensorInfo aclInput = armcomputetensorutils::BuildArmComputeTensorInfo(input);
22     const arm_compute::TensorInfo aclAlpha = armcomputetensorutils::BuildArmComputeTensorInfo(alpha);
23     const arm_compute::TensorInfo aclOutput = armcomputetensorutils::BuildArmComputeTensorInfo(output);
24 
25     return arm_compute::NEPReluLayer::validate(&aclInput,
26                                                &aclAlpha,
27                                                &aclOutput);
28 }
29 
NeonPreluWorkload(const PreluQueueDescriptor & descriptor,const WorkloadInfo & info)30 NeonPreluWorkload::NeonPreluWorkload(const PreluQueueDescriptor& descriptor,
31                                      const WorkloadInfo& info)
32         : NeonBaseWorkload<PreluQueueDescriptor>(descriptor, info)
33 {
34     m_Data.ValidateInputsOutputs("NeonPreluWorkload", 1, 1);
35 
36     arm_compute::ITensor& input = PolymorphicDowncast<IAclTensorHandle*>(m_Data.m_Inputs[0])->GetTensor();
37     arm_compute::ITensor& alpha = PolymorphicDowncast<IAclTensorHandle*>(m_Data.m_Inputs[1])->GetTensor();
38     arm_compute::ITensor& output = PolymorphicDowncast<IAclTensorHandle*>(m_Data.m_Outputs[0])->GetTensor();
39 
40     auto layer = std::make_unique<arm_compute::NEPReluLayer>();
41     layer->configure(&input, &alpha, &output);
42 
43     m_PreluLayer.reset(layer.release());
44 }
45 
Execute() const46 void NeonPreluWorkload::Execute() const
47 {
48     ARMNN_SCOPED_PROFILING_EVENT_NEON_GUID("NeonPreluWorkload_Execute", this->GetGuid());
49     m_PreluLayer->run();
50 }
51 
52 } //namespace armnn