Software Development
Optimized Launch for Multiple GPU Functions
CUDALink and OpenCLLink are optimized to run multiple GPU functions, for example, deciding when to keep memory on the GPU.
| In[1]:= |  X |
| Out[1]= |  |
| New in Wolfram Mathematica 8: CUDA and OpenCL Support | ◄ previous | next ► |
| In[1]:= | X |
| Out[1]= | |