Configure L1 cache / shared memory size?

3 views (last 30 days)
Hello, i am testing a ptx compiled kernel in MATLAB, card gtx 560 ti, cuda architecture 2.1. As the cuda programming guide writes:
"The same on-chip memory is used for both L1 and shared memory: It can be configured as 48 KB of shared memory and 16 KB of L1 cache or as 16 KB of shared memory and 48 KB of L1 cache, using cudaFuncSetCacheConfig()/cuFuncSetCacheConfig()"
i cant find any information on whether i can do this configuration in MATLAB, or with some option to nvcc when compiling the ptx.
The reason why i want to do this, is that my kernel doesnot use any shared memory, and i want to try whether i get any speedup with a bigger L1 cache. (the whole kernel calculates and reads double precision data)
Thanks for any solution!

Accepted Answer

Edric Ellis
Edric Ellis on 23 Jan 2014
You should be able to configure this by using a CUDA Mex file rather than CUDAKernel.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by