Right but my point is ... Right but my point is that I don 't think OpenCL code executes fast enough to tessellate thousands of tiles within a rendering pipeline. It 's not a matter of how quick the GPU processes the information once it 's got it, it 's that CUDA in C operates at ... |