gpu_info.cc 7.0 KB
Newer Older
1
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
L
liaogang 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14

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. */

Y
Yi Wang 已提交
15
#include "paddle/fluid/platform/gpu_info.h"
L
liaogang 已提交
16

17
#include <algorithm>
L
liaogang 已提交
18

19
#include "gflags/gflags.h"
Y
Yi Wang 已提交
20
#include "paddle/fluid/platform/enforce.h"
L
liaogang 已提交
21

22
DEFINE_double(fraction_of_gpu_memory_to_use, 0.92,
X
Xin Pan 已提交
23 24 25 26 27
              "Allocate a trunk of gpu memory that is this fraction of the "
              "total gpu memory size. Future memory usage will be allocated "
              "from the trunk. If the trunk doesn't have enough gpu memory, "
              "additional trunks of the same size will be requested from gpu "
              "until the gpu has no memory left for another trunk.");
L
liaogang 已提交
28

29 30 31 32 33 34 35 36 37 38
DEFINE_bool(
    enable_cublas_tensor_op_math, false,
    "The enable_cublas_tensor_op_math indicate whether to use Tensor Core, "
    "but it may loss precision. Currently, There are two CUDA libraries that"
    " use Tensor Cores, cuBLAS and cuDNN. cuBLAS uses Tensor Cores to speed up"
    " GEMM computations(the matrices must be either half precision or single "
    "precision); cuDNN uses Tensor Cores to speed up both convolutions(the "
    "input and output must be half precision) and recurrent neural networks "
    "(RNNs).");

L
liaogang 已提交
39 40 41
namespace paddle {
namespace platform {

42
int GetCUDADeviceCount() {
L
liaogang 已提交
43
  int count;
L
liaogang 已提交
44
  PADDLE_ENFORCE(
L
liaogang 已提交
45
      cudaGetDeviceCount(&count),
46
      "cudaGetDeviceCount failed in paddle::platform::GetCUDADeviceCount");
L
liaogang 已提交
47 48 49
  return count;
}

50 51 52 53 54 55 56 57 58
int GetCUDAComputeCapability(int id) {
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
  cudaDeviceProp device_prop;
  PADDLE_ENFORCE(cudaGetDeviceProperties(&device_prop, id),
                 "cudaGetDeviceProperties failed in "
                 "paddle::platform::GetCUDAComputeCapability");
  return device_prop.major * 10 + device_prop.minor;
}

C
chengduo 已提交
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
int GetCUDARuntimeVersion(int id) {
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
  int runtime_version = 0;
  PADDLE_ENFORCE(cudaRuntimeGetVersion(&runtime_version),
                 "cudaRuntimeGetVersion failed in "
                 "paddle::platform::cudaRuntimeGetVersion");
  return runtime_version;
}

int GetCUDADriverVersion(int id) {
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
  int driver_version = 0;
  PADDLE_ENFORCE(cudaDriverGetVersion(&driver_version),
                 "cudaDriverGetVersion failed in "
                 "paddle::platform::GetCUDADriverVersion");
  return driver_version;
}

77 78 79 80 81 82 83 84 85 86
bool TensorCoreAvailable() {
#if CUDA_VERSION >= 9000
  int device = GetCurrentDeviceId();
  int driver_version = GetCUDAComputeCapability(device);
  return driver_version >= 70;
#else
  return false;
#endif
}

C
chengduoZH 已提交
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
int GetCUDAMultiProcessors(int id) {
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
  int count;
  PADDLE_ENFORCE(
      cudaDeviceGetAttribute(&count, cudaDevAttrMultiProcessorCount, id),
      "cudaDeviceGetAttribute failed in "
      "paddle::platform::GetCUDAMultiProcessors");
  return count;
}

int GetCUDAMaxThreadsPerMultiProcessor(int id) {
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
  int count;
  PADDLE_ENFORCE(cudaDeviceGetAttribute(
                     &count, cudaDevAttrMaxThreadsPerMultiProcessor, id),
                 "cudaDeviceGetAttribute failed in "
                 "paddle::platform::GetCUDAMaxThreadsPerMultiProcessor");
  return count;
}

L
liaogang 已提交
107 108
int GetCurrentDeviceId() {
  int device_id;
L
liaogang 已提交
109
  PADDLE_ENFORCE(
L
liaogang 已提交
110 111 112 113 114 115
      cudaGetDevice(&device_id),
      "cudaGetDevice failed in paddle::platform::GetCurrentDeviceId");
  return device_id;
}

void SetDeviceId(int id) {
Q
qijun 已提交
116
  // TODO(qijun): find a better way to cache the cuda device count
Y
Yang Yang 已提交
117
  PADDLE_ENFORCE_LT(id, GetCUDADeviceCount(), "id must less than GPU count");
L
liaogang 已提交
118
  PADDLE_ENFORCE(cudaSetDevice(id),
L
liaogang 已提交
119 120 121
                 "cudaSetDevice failed in paddle::platform::SetDeviceId");
}

122 123
void GpuMemoryUsage(size_t *available, size_t *total) {
  PADDLE_ENFORCE(cudaMemGetInfo(available, total),
L
liaogang 已提交
124 125 126 127 128 129 130
                 "cudaMemGetInfo failed in paddle::platform::GetMemoryUsage");
}

size_t GpuMaxAllocSize() {
  size_t total = 0;
  size_t available = 0;

131
  GpuMemoryUsage(&available, &total);
L
liaogang 已提交
132

L
liaogang 已提交
133
  // Reserve the rest for page tables, etc.
L
liaogang 已提交
134
  return static_cast<size_t>(total * FLAGS_fraction_of_gpu_memory_to_use);
L
liaogang 已提交
135 136
}

L
liaogang 已提交
137 138 139 140 141 142 143
size_t GpuMinChunkSize() {
  // Allow to allocate the minimum chunk size is 256 bytes.
  return 1 << 8;
}

size_t GpuMaxChunkSize() {
  size_t total = 0;
C
chenweihang 已提交
144
  size_t available = 0;
L
liaogang 已提交
145

C
chenweihang 已提交
146
  GpuMemoryUsage(&available, &total);
147 148
  VLOG(100) << "GPU Usage " << available / 1024 / 1024 << "M/"
            << total / 1024 / 1024 << "M";
149
  size_t reserving = static_cast<size_t>(0.05 * total);
L
liaogang 已提交
150
  // If available less than minimum chunk size, no usable memory exists.
C
chenweihang 已提交
151 152 153
  available =
      std::min(std::max(available, GpuMinChunkSize()) - GpuMinChunkSize(),
               total - reserving);
154 155

  // Reserving the rest memory for page tables, etc.
L
liaogang 已提交
156

C
chenweihang 已提交
157 158
  size_t allocating = static_cast<size_t>(FLAGS_fraction_of_gpu_memory_to_use *
                                          (total - reserving));
L
liaogang 已提交
159

C
chenweihang 已提交
160 161
  PADDLE_ENFORCE_LE(allocating, available,
                    "Insufficient GPU memory to allocation.");
162

C
chenweihang 已提交
163
  return allocating;
L
liaogang 已提交
164 165
}

L
liaogang 已提交
166 167
void GpuMemcpyAsync(void *dst, const void *src, size_t count,
                    enum cudaMemcpyKind kind, cudaStream_t stream) {
L
liaogang 已提交
168 169
  PADDLE_ENFORCE(cudaMemcpyAsync(dst, src, count, kind, stream),
                 "cudaMemcpyAsync failed in paddle::platform::GpuMemcpyAsync");
L
liaogang 已提交
170 171
}

172 173 174 175 176 177 178 179
void GpuMemcpySync(void *dst, const void *src, size_t count,
                   enum cudaMemcpyKind kind) {
  PADDLE_ENFORCE(cudaMemcpy(dst, src, count, kind),
                 "cudaMemcpy failed in paddle::platform::GpuMemcpySync");
}

void GpuMemcpyPeerAsync(void *dst, int dst_device, const void *src,
                        int src_device, size_t count, cudaStream_t stream) {
L
liaogang 已提交
180
  PADDLE_ENFORCE(
L
liaogang 已提交
181
      cudaMemcpyPeerAsync(dst, dst_device, src, src_device, count, stream),
182 183 184 185 186 187 188 189
      "cudaMemcpyPeerAsync failed in paddle::platform::GpuMemcpyPeerAsync");
}

void GpuMemcpyPeerSync(void *dst, int dst_device, const void *src,
                       int src_device, size_t count) {
  PADDLE_ENFORCE(
      cudaMemcpyPeer(dst, dst_device, src, src_device, count),
      "cudaMemcpyPeer failed in paddle::platform::GpuMemcpyPeerSync");
L
liaogang 已提交
190
}
D
dzhwinter 已提交
191 192 193 194 195

void GpuMemsetAsync(void *dst, int value, size_t count, cudaStream_t stream) {
  PADDLE_ENFORCE(cudaMemsetAsync(dst, value, count, stream),
                 "cudaMemsetAsync failed in paddle::platform::GpuMemsetAsync");
}
L
liaogang 已提交
196 197
}  // namespace platform
}  // namespace paddle