Relative Content

Tag Archive for c++parallel-processingcudagpumcmc

How to use persistent kernel (using cooperative groups) for the following kernel structure?

I have a kernel which is supposed to generate statistics on a specific simulation by generating samples. I need a (very, very) large number of samples (possibly needing kernel to run for hours). I have allocated a working space (an array) for a number of samples N (which is less than the total number of samples) and reuse space by resetting the the array index to a seed.