デバイス情報の取得
をテンプレートにして作成
[
トップ
|
新規
|
一覧
|
検索
|
最終更新
|
ヘルプ
]
開始行:
*デバイスプロパティの取得 [#e52f7f5b]
デバイスの情報は cudaGetDeviceProperties命令でデバイスプ...
#code(C){{
int n; //デバイス数
cutilSafeCall(cudaGetDeviceCount(&n));
for(int i = 0; i < n; ++i){
cudaDeviceProp dev;
// デバイスプロパティ取得
cutilSafeCall(cudaGetDeviceProperties(&dev, i));
printf("device %d\n", i);
printf(" device name : %s\n", dev.name);
printf(" total global memory : %d (MB)\n", dev.totalGlo...
printf(" shared memory / block : %d (KB)\n", dev.shared...
printf(" register / block : %d\n", dev.regsPerBlock);
printf(" warp size : %d\n", dev.warpSize);
printf(" max pitch : %d (B)\n", dev.memPitch);
printf(" max threads / block : %d\n", dev.maxThreadsPer...
printf(" max size of each dim. of block : (%d, %d, %d)\...
printf(" max size of each dim. of grid : (%d, %d, %d)\...
printf(" clock rate : %d (MHz)\n", dev.clockRate/1000);
printf(" total constant memory : %d (KB)\n", dev.totalC...
printf(" compute capability : %d.%d\n", dev.major, dev....
printf(" alignment requirement for texture : %d\n", dev...
printf(" device overlap : %s\n", (dev.deviceOverlap ? "...
printf(" num. of multiprocessors : %d\n", dev.multiProc...
printf(" kernel execution timeout : %s\n", (dev.kernelE...
printf(" integrated : %s\n", (dev.integrated ? "on" : "...
printf(" host memory mapping : %s\n", (dev.canMapHostMe...
printf(" compute mode : ");
if(dev.computeMode == cudaComputeModeDefault) printf("d...
else if(dev.computeMode == cudaComputeModeExclusive) pr...
else if(dev.computeMode == cudaComputeModeProhibited) p...
}
}}
***GeForce GTX 285を搭載したビデオカード上での実行例 [#x0...
device 0
device name : GeForce GTX 285
total global memory : 1024 (MB)
shared memory / block : 16 (KB)
register / block : 16384
warp size : 32
max pitch : 262144 (B)
max threads / block : 512
max size of each dim. of block : (512, 512, 64)
max size of each dim. of grid : (65535, 65535, 1)
clock rate : 1476 (MHz)
total constant memory : 64 (KB)
compute capability : 1.3
alignment requirement for texture : 256
device overlap : ok
num. of multiprocessors : 30
kernel execution timeout : off
integrated : off
host memory mapping : off
compute mode : default mode (multiple threads can use)
***GeForce GTX 580を搭載したビデオカード上での実行例 [#jf...
device 0
device name : GeForce GTX 580
total global memory : 1503 (MB)
shared memory / block : 48 (KB)
register / block : 32768
warp size : 32
max pitch : 2147483647 (B)
max threads / block : 1024
max size of each dim. of block : (1024, 1024, 64)
max size of each dim. of grid : (65535, 65535, 65535)
clock rate : 1544 (MHz)
total constant memory : 64 (KB)
compute capability : 2.0
alignment requirement for texture : 512
device overlap : ok
num. of multiprocessors : 16
kernel execution timeout : on
integrated : off
host memory mapping : on
compute mode : default mode (multiple threads can use)
***GeForce GTX TITANを搭載したビデオカード上での実行例 [#...
device 0
device name : GeForce GTX TITAN
total global memory : 4095 (MB)
shared memory / block : 48 (KB)
register / block : 65536
warp size : 32
max pitch : 2147483647 (B)
max threads / block : 1024
max size of each dim. of block : (1024, 1024, 64)
max size of each dim. of grid : (2147483647, 65535, 65...
clock rate : 875 (MHz)
total constant memory : 64 (KB)
compute capability : 3.5
alignment requirement for texture : 512
device overlap : ok
num. of multiprocessors : 14
kernel execution timeout : on
integrated : off
host memory mapping : on
compute mode : default mode (multiple threads can use)
終了行:
*デバイスプロパティの取得 [#e52f7f5b]
デバイスの情報は cudaGetDeviceProperties命令でデバイスプ...
#code(C){{
int n; //デバイス数
cutilSafeCall(cudaGetDeviceCount(&n));
for(int i = 0; i < n; ++i){
cudaDeviceProp dev;
// デバイスプロパティ取得
cutilSafeCall(cudaGetDeviceProperties(&dev, i));
printf("device %d\n", i);
printf(" device name : %s\n", dev.name);
printf(" total global memory : %d (MB)\n", dev.totalGlo...
printf(" shared memory / block : %d (KB)\n", dev.shared...
printf(" register / block : %d\n", dev.regsPerBlock);
printf(" warp size : %d\n", dev.warpSize);
printf(" max pitch : %d (B)\n", dev.memPitch);
printf(" max threads / block : %d\n", dev.maxThreadsPer...
printf(" max size of each dim. of block : (%d, %d, %d)\...
printf(" max size of each dim. of grid : (%d, %d, %d)\...
printf(" clock rate : %d (MHz)\n", dev.clockRate/1000);
printf(" total constant memory : %d (KB)\n", dev.totalC...
printf(" compute capability : %d.%d\n", dev.major, dev....
printf(" alignment requirement for texture : %d\n", dev...
printf(" device overlap : %s\n", (dev.deviceOverlap ? "...
printf(" num. of multiprocessors : %d\n", dev.multiProc...
printf(" kernel execution timeout : %s\n", (dev.kernelE...
printf(" integrated : %s\n", (dev.integrated ? "on" : "...
printf(" host memory mapping : %s\n", (dev.canMapHostMe...
printf(" compute mode : ");
if(dev.computeMode == cudaComputeModeDefault) printf("d...
else if(dev.computeMode == cudaComputeModeExclusive) pr...
else if(dev.computeMode == cudaComputeModeProhibited) p...
}
}}
***GeForce GTX 285を搭載したビデオカード上での実行例 [#x0...
device 0
device name : GeForce GTX 285
total global memory : 1024 (MB)
shared memory / block : 16 (KB)
register / block : 16384
warp size : 32
max pitch : 262144 (B)
max threads / block : 512
max size of each dim. of block : (512, 512, 64)
max size of each dim. of grid : (65535, 65535, 1)
clock rate : 1476 (MHz)
total constant memory : 64 (KB)
compute capability : 1.3
alignment requirement for texture : 256
device overlap : ok
num. of multiprocessors : 30
kernel execution timeout : off
integrated : off
host memory mapping : off
compute mode : default mode (multiple threads can use)
***GeForce GTX 580を搭載したビデオカード上での実行例 [#jf...
device 0
device name : GeForce GTX 580
total global memory : 1503 (MB)
shared memory / block : 48 (KB)
register / block : 32768
warp size : 32
max pitch : 2147483647 (B)
max threads / block : 1024
max size of each dim. of block : (1024, 1024, 64)
max size of each dim. of grid : (65535, 65535, 65535)
clock rate : 1544 (MHz)
total constant memory : 64 (KB)
compute capability : 2.0
alignment requirement for texture : 512
device overlap : ok
num. of multiprocessors : 16
kernel execution timeout : on
integrated : off
host memory mapping : on
compute mode : default mode (multiple threads can use)
***GeForce GTX TITANを搭載したビデオカード上での実行例 [#...
device 0
device name : GeForce GTX TITAN
total global memory : 4095 (MB)
shared memory / block : 48 (KB)
register / block : 65536
warp size : 32
max pitch : 2147483647 (B)
max threads / block : 1024
max size of each dim. of block : (1024, 1024, 64)
max size of each dim. of grid : (2147483647, 65535, 65...
clock rate : 875 (MHz)
total constant memory : 64 (KB)
compute capability : 3.5
alignment requirement for texture : 512
device overlap : ok
num. of multiprocessors : 14
kernel execution timeout : on
integrated : off
host memory mapping : on
compute mode : default mode (multiple threads can use)
ページ名:
![[PukiWiki]](image/pukiwiki.png "[PukiWiki]")
