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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
// Copyright 2020-2022, NVIDIA CORPORATION. All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions
// are met:
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
// * Neither the name of NVIDIA CORPORATION nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
// OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "backend_memory_manager.h"
#include "pinned_memory_manager.h"
#include "status.h"
#include "tritonserver_apis.h"
#ifdef TRITON_ENABLE_GPU
#include <cuda_runtime_api.h>
#include "cuda_memory_manager.h"
#endif // TRITON_ENABLE_GPU
// For unknown reason, windows will not export the TRITONBACKEND_*
// functions declared with dllexport in tritonbackend.h. To get those
// functions exported it is (also?) necessary to mark the definitions
// in this file with dllexport as well.
#if defined(_MSC_VER)
#define TRITONAPI_DECLSPEC __declspec(dllexport)
#elif defined(__GNUC__)
#define TRITONAPI_DECLSPEC __attribute__((__visibility__("default")))
#else
#define TRITONAPI_DECLSPEC
#endif
namespace triton { namespace core {
extern "C" {
TRITONAPI_DECLSPEC TRITONSERVER_Error*
TRITONBACKEND_MemoryManagerAllocate(
TRITONBACKEND_MemoryManager* manager, void** buffer,
const TRITONSERVER_MemoryType memory_type, const int64_t memory_type_id,
const uint64_t byte_size)
{
switch (memory_type) {
case TRITONSERVER_MEMORY_GPU:
#ifdef TRITON_ENABLE_GPU
{
auto status = CudaMemoryManager::Alloc(buffer, byte_size, memory_type_id);
if (!status.IsOk()) {
return TRITONSERVER_ErrorNew(
StatusCodeToTritonCode(status.ErrorCode()),
status.Message().c_str());
}
break;
}
#else
return TRITONSERVER_ErrorNew(
TRITONSERVER_ERROR_UNSUPPORTED,
"GPU memory allocation not supported");
#endif // TRITON_ENABLE_GPU
case TRITONSERVER_MEMORY_CPU_PINNED:
#ifdef TRITON_ENABLE_GPU
{
TRITONSERVER_MemoryType mt = memory_type;
auto status = PinnedMemoryManager::Alloc(buffer, byte_size, &mt, false);
if (!status.IsOk()) {
return TRITONSERVER_ErrorNew(
StatusCodeToTritonCode(status.ErrorCode()),
status.Message().c_str());
}
break;
}
#else
return TRITONSERVER_ErrorNew(
TRITONSERVER_ERROR_UNSUPPORTED,
"Pinned memory allocation not supported");
#endif // TRITON_ENABLE_GPU
case TRITONSERVER_MEMORY_CPU: {
*buffer = malloc(byte_size);
if (*buffer == nullptr) {
return TRITONSERVER_ErrorNew(
TRITONSERVER_ERROR_UNAVAILABLE, "CPU memory allocation failed");
}
break;
}
}
return nullptr; // success
}
TRITONAPI_DECLSPEC TRITONSERVER_Error*
TRITONBACKEND_MemoryManagerFree(
TRITONBACKEND_MemoryManager* manager, void* buffer,
const TRITONSERVER_MemoryType memory_type, const int64_t memory_type_id)
{
switch (memory_type) {
case TRITONSERVER_MEMORY_GPU: {
#ifdef TRITON_ENABLE_GPU
auto status = CudaMemoryManager::Free(buffer, memory_type_id);
if (!status.IsOk()) {
return TRITONSERVER_ErrorNew(
StatusCodeToTritonCode(status.StatusCode()),
status.Message().c_str());
}
#endif // TRITON_ENABLE_GPU
break;
}
case TRITONSERVER_MEMORY_CPU_PINNED: {
#ifdef TRITON_ENABLE_GPU
auto status = PinnedMemoryManager::Free(buffer);
if (!status.IsOk()) {
return TRITONSERVER_ErrorNew(
StatusCodeToTritonCode(status.StatusCode()),
status.Message().c_str());
}
#endif // TRITON_ENABLE_GPU
break;
}
case TRITONSERVER_MEMORY_CPU:
free(buffer);
break;
}
return nullptr; // success
}
} // extern C
}} // namespace triton::core