blob: 4992e9c07a18052b6e4e4cd8ac3655d0fe1818da (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
|
//
// Copyright © 2022 Arm Ltd and Contributors. All rights reserved.
// SPDX-License-Identifier: MIT
//
#include "RefConvertFp32ToFp16Workload.hpp"
#include "RefWorkloadUtils.hpp"
#include "Profiling.hpp"
#include <armnnUtils/FloatingPointConverter.hpp>
#include <Half.hpp>
namespace armnn
{
void RefConvertFp32ToFp16Workload::Execute() const
{
Execute(m_Data.m_Inputs, m_Data.m_Outputs);
}
void RefConvertFp32ToFp16Workload::ExecuteAsync(ExecutionData& executionData)
{
WorkingMemDescriptor* workingMemDescriptor = static_cast<WorkingMemDescriptor*>(executionData.m_Data);
Execute(workingMemDescriptor->m_Inputs, workingMemDescriptor->m_Outputs);
}
void RefConvertFp32ToFp16Workload::Execute(std::vector<ITensorHandle*> inputs,
std::vector<ITensorHandle*> outputs) const
{
ARMNN_SCOPED_PROFILING_EVENT(Compute::CpuRef, "RefConvertFp32ToFp16Workload_Execute");
const float* const input = reinterpret_cast<const float*>(inputs[0]->Map());
Half* const output = reinterpret_cast<Half*>(outputs[0]->Map());
// convert Fp32 input to Fp16 output
unsigned int numElements = GetTensorInfo(inputs[0]).GetNumElements();
armnnUtils::FloatingPointConverter::ConvertFloat32To16(input, numElements, output);
}
} //namespace armnn
|
