From 6940dd720ebb6b3d1df8ca203ab696daefe58189 Mon Sep 17 00:00:00 2001 From: Jim Flynn Date: Fri, 20 Mar 2020 12:25:56 +0000 Subject: renamed Documentation folder 20.02 and added .nojekyll file Signed-off-by: Jim Flynn --- ...convert_fp32_to_fp16_workload_8cpp_source.xhtml | 131 +++++++++++++++++++++ 1 file changed, 131 insertions(+) create mode 100644 20.02/_neon_convert_fp32_to_fp16_workload_8cpp_source.xhtml (limited to '20.02/_neon_convert_fp32_to_fp16_workload_8cpp_source.xhtml') diff --git a/20.02/_neon_convert_fp32_to_fp16_workload_8cpp_source.xhtml b/20.02/_neon_convert_fp32_to_fp16_workload_8cpp_source.xhtml new file mode 100644 index 0000000000..29c7de31fc --- /dev/null +++ b/20.02/_neon_convert_fp32_to_fp16_workload_8cpp_source.xhtml @@ -0,0 +1,131 @@ + + + + + + + + + + + + + +ArmNN: src/backends/neon/workloads/NeonConvertFp32ToFp16Workload.cpp Source File + + + + + + + + + + + + + + + + +
+
+ + + + ArmNN + + + +
+
+  20.02 +
+
+
+ + + + + + + +
+
+ +
+
+
+ +
+ +
+
+ + +
+ +
+ +
+
+
NeonConvertFp32ToFp16Workload.cpp
+
+
+Go to the documentation of this file.
1 //
2 // Copyright © 2017 Arm Ltd. All rights reserved.
3 // SPDX-License-Identifier: MIT
4 //
5 
6 #include "NeonConvertFp32ToFp16Workload.hpp"
7 
8 #include <Half.hpp>
9 #include <Profiling.hpp>
10 
11 #include <armnnUtils/FloatingPointConverter.hpp>
12 
13 #include <backendsCommon/WorkloadUtils.hpp>
14 
15 namespace armnn
16 {
17 
18 NeonConvertFp32ToFp16Workload::NeonConvertFp32ToFp16Workload(const ConvertFp32ToFp16QueueDescriptor& descriptor,
19  const WorkloadInfo& info)
20  : Float32ToFloat16Workload<ConvertFp32ToFp16QueueDescriptor>(descriptor, info)
21 {
22  this->m_Data.ValidateInputsOutputs("NeonConvertFp32ToFp16Workload", 1, 1);
23  GatherTensorHandlePairs(descriptor, m_TensorHandlePairs);
24 }
25 
26 void NeonConvertFp32ToFp16Workload::Execute() const
27 {
28  ARMNN_SCOPED_PROFILING_EVENT_NEON("NeonConvertFp32ToFp16Workload_Execute");
29 
30  auto convertFunc = [](uint8_t* dst, const uint8_t* src, size_t size)
31  {
32  auto input = reinterpret_cast<const float*>(src);
33  auto output = reinterpret_cast<Half*>(dst);
34  size_t numElements = size/2; // 2 bytes per fp16
35  armnnUtils::FloatingPointConverter::ConvertFloat32To16(input, numElements, output);
36  };
37 
38  for (const auto& pair : m_TensorHandlePairs)
39  {
40  CopyTensorContentsGeneric(pair.first, pair.second, convertFunc);
41  }
42 }
43 
44 } //namespace armnn
FloatingPointConverter.hpp
+ +
armnn::BaseWorkload::m_Data
const QueueDescriptor m_Data
Definition: Workload.hpp:46
+
ARMNN_SCOPED_PROFILING_EVENT_NEON
#define ARMNN_SCOPED_PROFILING_EVENT_NEON(name)
Definition: NeonWorkloadUtils.hpp:17
+
armnn::QueueDescriptor::ValidateInputsOutputs
void ValidateInputsOutputs(const std::string &descName, unsigned int numExpectedIn, unsigned int numExpectedOut) const
Definition: WorkloadData.cpp:461
+
armnn
Copyright (c) 2020 ARM Limited.
Definition: 00_introduction.dox:25
+ +
armnnUtils::FloatingPointConverter::ConvertFloat32To16
static void ConvertFloat32To16(const float *srcFloat32Buffer, size_t numElements, void *dstFloat16Buffer)
Converts a buffer of FP32 values to FP16, and stores in the given dstFloat16Buffer.
Definition: FloatingPointConverter.cpp:16
+ +
Half.hpp
+
armnn::NeonConvertFp32ToFp16Workload::NeonConvertFp32ToFp16Workload
NeonConvertFp32ToFp16Workload(const ConvertFp32ToFp16QueueDescriptor &descriptor, const WorkloadInfo &info)
Definition: NeonConvertFp32ToFp16Workload.cpp:18
+
armnn::WorkloadInfo
Contains information about inputs and outputs to a layer.
Definition: WorkloadInfo.hpp:16
+
NeonConvertFp32ToFp16Workload.hpp
+
armnn::CopyTensorContentsGeneric
void CopyTensorContentsGeneric(const ITensorHandle *srcTensor, ITensorHandle *dstTensor, CopyFunc copy)
Definition: WorkloadUtils.hpp:49
+
armnn::GatherTensorHandlePairs
void GatherTensorHandlePairs(const DescriptorType &descriptor, std::vector< std::pair< SrcTensorHandleType *, DstTensorHandleType *>> &tensorHandlePairs)
Definition: WorkloadUtils.hpp:192
+
armnn::Half
half_float::half Half
Definition: Half.hpp:16
+
Profiling.hpp
+
WorkloadUtils.hpp
+
+
+ + + + -- cgit v1.2.1