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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
|
#include <cuda_runtime.h>
#include <stdio.h>
#include <stdlib.h>
#ifndef PRINT_SUFFIX
#define PRINT_SUFFIX "<find_cudadevices>"
#endif
#define MY_CUDA_VER (__CUDACC_VER_MAJOR__ * 100 + __CUDACC_VER_MINOR__)
inline void check(cudaError_t result)
{
if (result)
{
fprintf(stderr, PRINT_SUFFIX "%s (%s)", cudaGetErrorName(result), cudaGetErrorString(result));
cudaDeviceReset();
// Make sure we call CUDA Device Reset before exiting
exit(0);
}
}
inline void print_value(unsigned long long value)
{
printf("%llu", value);
}
inline void print_value(unsigned long value)
{
printf("%lu", value);
}
inline void print_value(unsigned int value)
{
printf("%u", value);
}
inline void print_value(bool value)
{
printf(value ? "true" : "false");
}
inline void print_value(int value)
{
printf("%d", value);
}
template <typename T, size_t len>
inline void print_value(const T (&value)[len])
{
printf("(");
for (size_t i = 0; i < len - 1; i++)
{
print_value(value[i]);
printf(", ");
}
print_value(value[len - 1]);
printf(")");
}
inline void print_value(const void *value)
{
printf("\"%s\"", (const char *)value);
}
template <size_t len>
inline void print_value(const char (&value)[len])
{
printf("\"");
for (size_t i = 0; i < len; i++)
printf("%02hhx", value[i]);
printf("\"");
}
template <>
inline void print_value<16>(const char (&value)[16])
{
// speicalized for uuid
printf("\"%02hhx%02hhx%02hhx%02hhx-%02hhx%02hhx-%02hhx%02hhx-%02hhx%02hhx-%02hhx%02hhx%02hhx%02hhx%02hhx%02hhx\"",
value[0], value[1], value[2], value[3],
value[4], value[5], value[6], value[7],
value[8], value[9], value[10], value[11],
value[12], value[13], value[14], value[15]);
}
#if MY_CUDA_VER >= 1000
inline void print_value(const cudaUUID_t &value)
{
print_value(value.bytes);
}
#endif
template <typename T>
inline void print_property(const char *name, const T &value)
{
printf(PRINT_SUFFIX " %s = ", name);
print_value(value);
printf("\n");
}
inline void print_device(int id)
{
cudaDeviceProp deviceProp;
check(cudaGetDeviceProperties(&deviceProp, id));
#define PRINT_PROPERTY(name) print_property(#name, deviceProp.name)
#define PRINT_BOOL_PROPERTY(name) print_property(#name, static_cast<bool>(deviceProp.name))
#define PRINT_STR_PROPERTY(name) print_property(#name, static_cast<const void *>(deviceProp.name))
// cuda 8.0
PRINT_STR_PROPERTY(name);
PRINT_PROPERTY(totalGlobalMem);
PRINT_PROPERTY(sharedMemPerBlock);
PRINT_PROPERTY(regsPerBlock);
PRINT_PROPERTY(warpSize);
PRINT_PROPERTY(memPitch);
PRINT_PROPERTY(maxThreadsPerBlock);
PRINT_PROPERTY(maxThreadsDim);
PRINT_PROPERTY(maxGridSize);
PRINT_PROPERTY(clockRate);
PRINT_PROPERTY(totalConstMem);
PRINT_PROPERTY(major);
PRINT_PROPERTY(minor);
PRINT_PROPERTY(textureAlignment);
PRINT_PROPERTY(texturePitchAlignment);
PRINT_BOOL_PROPERTY(deviceOverlap);
PRINT_PROPERTY(multiProcessorCount);
PRINT_BOOL_PROPERTY(kernelExecTimeoutEnabled);
PRINT_BOOL_PROPERTY(integrated);
PRINT_BOOL_PROPERTY(canMapHostMemory);
PRINT_PROPERTY(computeMode);
PRINT_PROPERTY(maxTexture1D);
PRINT_PROPERTY(maxTexture1DMipmap);
PRINT_PROPERTY(maxTexture1DLinear);
PRINT_PROPERTY(maxTexture2D);
PRINT_PROPERTY(maxTexture2DMipmap);
PRINT_PROPERTY(maxTexture2DLinear);
PRINT_PROPERTY(maxTexture2DGather);
PRINT_PROPERTY(maxTexture3D);
PRINT_PROPERTY(maxTexture3DAlt);
PRINT_PROPERTY(maxTextureCubemap);
PRINT_PROPERTY(maxTexture1DLayered);
PRINT_PROPERTY(maxTexture2DLayered);
PRINT_PROPERTY(maxTextureCubemapLayered);
PRINT_PROPERTY(maxSurface1D);
PRINT_PROPERTY(maxSurface2D);
PRINT_PROPERTY(maxSurface3D);
PRINT_PROPERTY(maxSurface1DLayered);
PRINT_PROPERTY(maxSurface2DLayered);
PRINT_PROPERTY(maxSurfaceCubemap);
PRINT_PROPERTY(maxSurfaceCubemapLayered);
PRINT_PROPERTY(surfaceAlignment);
PRINT_BOOL_PROPERTY(concurrentKernels);
PRINT_BOOL_PROPERTY(ECCEnabled);
PRINT_PROPERTY(pciBusID);
PRINT_PROPERTY(pciDeviceID);
PRINT_PROPERTY(pciDomainID);
PRINT_BOOL_PROPERTY(tccDriver);
PRINT_PROPERTY(asyncEngineCount);
PRINT_BOOL_PROPERTY(unifiedAddressing);
PRINT_PROPERTY(memoryClockRate);
PRINT_PROPERTY(memoryBusWidth);
PRINT_PROPERTY(l2CacheSize);
PRINT_PROPERTY(maxThreadsPerMultiProcessor);
PRINT_BOOL_PROPERTY(streamPrioritiesSupported);
PRINT_BOOL_PROPERTY(globalL1CacheSupported);
PRINT_BOOL_PROPERTY(localL1CacheSupported);
PRINT_PROPERTY(sharedMemPerMultiprocessor);
PRINT_PROPERTY(regsPerMultiprocessor);
PRINT_BOOL_PROPERTY(isMultiGpuBoard);
PRINT_PROPERTY(multiGpuBoardGroupID);
PRINT_PROPERTY(singleToDoublePrecisionPerfRatio);
PRINT_BOOL_PROPERTY(pageableMemoryAccess);
PRINT_BOOL_PROPERTY(concurrentManagedAccess);
PRINT_BOOL_PROPERTY(managedMemory);
#if MY_CUDA_VER >= 900
// Added in cuda 9.0
PRINT_BOOL_PROPERTY(computePreemptionSupported);
PRINT_BOOL_PROPERTY(canUseHostPointerForRegisteredMem);
PRINT_BOOL_PROPERTY(cooperativeLaunch);
PRINT_BOOL_PROPERTY(cooperativeMultiDeviceLaunch);
PRINT_PROPERTY(sharedMemPerBlockOptin);
#endif
#if MY_CUDA_VER >= 902
// Added in cuda 9.2
PRINT_BOOL_PROPERTY(pageableMemoryAccessUsesHostPageTables);
PRINT_BOOL_PROPERTY(directManagedMemAccessFromHost);
#endif
#if MY_CUDA_VER >= 1000
// Added in cuda 10.0
PRINT_PROPERTY(uuid);
PRINT_PROPERTY(luid);
PRINT_PROPERTY(luidDeviceNodeMask);
#endif
}
int main(int argc, char *argv[])
{
printf("\n");
fprintf(stderr, "\n");
int count = 0;
check(cudaGetDeviceCount(&count));
for (int i = 0; i < count; i++)
{
printf(PRINT_SUFFIX "DEVICE #%d\n", i);
print_device(i);
}
return 0;
}
|
