MATLAB PARALLEL COMPUTING TOOLBOX - S Guida Utente Scaricare il pdf (Pagina 290)

100

101

9 GPU Computing

9-32

{

int i = blockDim.x * blockIdx.x + threadIdx.x;

if (i < N)

B[i] = 2.0 * A[i];

}

It contains the following lines to determine the array size and launch a grid of the proper

size:

N = (int)(mxGPUGetNumberOfElements(A));

blocksPerGrid = (N + threadsPerBlock - 1) / threadsPerBlock;

TimesTwo<<<blocksPerGrid, threadsPerBlock>>>(d_A, d_B, N);

Set Up for MEX-File Compilation

• Your MEX source file that includes CUDA code must have a name with the extension

.cu, not .c nor .cpp.

• Before you compile your MEX-file, copy the provided mex_CUDA_platform.xml file

for your platform and compiler from the specified location, into the same folder as

your MEX source file.

•

For Linux

platforms:

matlabroot/toolbox/distcomp/gpu/extern/src/mex/glnxa64/gcc/mex_CUDA_glnxa64.xml

• For Macintosh platforms:

matlabroot/toolbox/distcomp/gpu/extern/src/mex/maci64/clang/mex_CUDA_maci64.xml

•

For Windows platforms with Visual Studio

2012:

matlabroot\toolbox\distcomp\gpu\extern\src\mex\win64\msvc2012\mex_CUDA_win64.xml

• For Windows platforms with Visual Studio 2013:

matlabroot\toolbox\distcomp\gpu\extern\src\mex\win64\msvc2013\mex_CUDA_win64.xml

• You must use the version of the NVIDIA compiler (nvcc) consistent with the

ToolkitVersion property of the GPUDevice object.

• Before compiling, make sure either that the location of the folder containing nvcc

is on your search path, or that the path to this folder is encoded in the environment

variable MW_NVCC_PATH. You can set this variable using the MATLAB setenv

command. For example:

setenv('MW_NVCC_PATH','/usr/local/CUDA/bin')

1 2 ... 285 286 287 288 289 290 291 292 293 294 295 ... 655 656

Commenti su questo manuale

Nessun commento

MATLAB PARALLEL COMPUTING TOOLBOX - S Guida Utente Pagina 290

Commenti su questo manuale

Prodotti e manuali riguardandi Software MATLAB PARALLEL COMPUTING TOOLBOX - S