blob: 39ae14eaf3f472b7652b15b20918333ca72f1f01 (
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
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
|
//
// Copyright © 2017 Arm Ltd. All rights reserved.
// SPDX-License-Identifier: MIT
//
#include "ClConstantWorkload.hpp"
#include <Half.hpp>
#include <aclCommon/ArmComputeTensorUtils.hpp>
#include <cl/ClTensorHandle.hpp>
#include <backendsCommon/CpuTensorHandle.hpp>
#include "ClWorkloadUtils.hpp"
namespace armnn
{
ClConstantWorkload::ClConstantWorkload(const ConstantQueueDescriptor& descriptor, const WorkloadInfo& info)
: BaseWorkload<ConstantQueueDescriptor>(descriptor, info)
, m_RanOnce(false)
{
}
void ClConstantWorkload::Execute() const
{
ARMNN_SCOPED_PROFILING_EVENT_CL("ClConstantWorkload_Execute");
// The intermediate tensor held by the corresponding layer output handler can be initialised with the given data
// on the first inference, then reused for subsequent inferences.
// The initialisation cannot happen at workload construction time since the ACL kernel for the next layer may not
// have been configured at the time.
if (!m_RanOnce)
{
const ConstantQueueDescriptor& data = this->m_Data;
BOOST_ASSERT(data.m_LayerOutput != nullptr);
arm_compute::CLTensor& output = static_cast<ClTensorHandle*>(data.m_Outputs[0])->GetTensor();
arm_compute::DataType computeDataType = static_cast<ClTensorHandle*>(data.m_Outputs[0])->GetDataType();
switch (computeDataType)
{
case arm_compute::DataType::F16:
{
CopyArmComputeClTensorData(output, data.m_LayerOutput->GetConstTensor<Half>());
break;
}
case arm_compute::DataType::F32:
{
CopyArmComputeClTensorData(output, data.m_LayerOutput->GetConstTensor<float>());
break;
}
case arm_compute::DataType::QASYMM8:
{
CopyArmComputeClTensorData(output, data.m_LayerOutput->GetConstTensor<uint8_t>());
break;
}
default:
{
BOOST_ASSERT_MSG(false, "Unknown data type");
break;
}
}
m_RanOnce = true;
}
}
} //namespace armnn
|
