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
69
70
71
72
73
74
75
76
77
78
79
80
81
|
/*
* Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <arm_compute/runtime/CL/CLScheduler.h>
#include <arm_compute/runtime/CL/CLMemory.h>
#include <arm_compute/runtime/CL/CLMemoryRegion.h>
#include "CLTensor.h"
#include "backend/acl_cl/Convert.h"
namespace neurun
{
namespace backend
{
namespace acl_cl
{
namespace operand
{
CLTensor::CLTensor(const compiler::TensorInfo &info)
: _cl_tensor(std::make_shared<arm_compute::CLTensor>())
{
auto acl_cl_info = asTensorInfo(info.shape(), info.typeInfo());
allocator()->init(acl_cl_info);
}
arm_compute::CLTensor *CLTensor::handle() const { return _cl_tensor.get(); }
arm_compute::CLTensor *CLTensor::handle() { return _cl_tensor.get(); }
arm_compute::CLTensorAllocator *CLTensor::allocator() { return _cl_tensor->allocator(); }
void CLTensor::map(bool blocking) { _cl_tensor->map(blocking); }
void CLTensor::unmap() { _cl_tensor->unmap(); }
uint8_t *CLTensor::doMap(cl::CommandQueue &q, bool blocking)
{
return allocator()->map(q, blocking);
}
void CLTensor::doUnmap(cl::CommandQueue &q) { allocator()->unmap(q, buffer()); }
// handle() is Deprecated on acl v18.11
// TODO Update this
#if 0
void CLTensor::setBuffer(void *host_ptr)
{
// create empty MemoryRegion: just context. Since flag isn't used here, no matter which flag to
// pass
auto memory = arm_compute::CLMemory(std::make_shared<arm_compute::CLBufferMemoryRegion>(
arm_compute::CLScheduler::get().context(), CL_MEM_USE_HOST_PTR | CL_MEM_READ_WRITE, 0));
// set buffer
auto mem = reinterpret_cast<cl::Buffer *>(memory.region()->handle());
*mem = cl::Buffer(arm_compute::CLScheduler::get().context(),
CL_MEM_USE_HOST_PTR | CL_MEM_READ_WRITE, info()->total_size(), host_ptr);
// set correct buffer size
memory.region()->set_size(info()->total_size());
// import memory
allocator()->import_memory(memory);
}
#endif
} // namespace operand
} // namespace acl_cl
} // namespace backend
} // namespace neurun
|