PDA

View Full Version : L1 cache and system memory split



driver
11-17-2014, 10:11 AM
Hi

Under Cuda, on a per-kernel basis, one can split the local memory (64KB) on a fermi in a 16KB/48KB or a 48KB/16KB format at runtime through the use of
cudaFuncSetCacheConfig() function.

Is there a way I can set something similar to this through OpenGL on a function/shader call?

thanks