eyalroz / gpu-kernel-runner

Runs a single CUDA/OpenCL kernel, taking its source from a file and arguments from the command-line

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Support flushing the L2 cache before each invocation

eyalroz opened this issue · comments

At the moment, the kernel invocations begins with an caches which may well contain much of the data used in the previous invocation. While this also happens in real life sometimes - it is often better to time kernel execution after flushing the caches.

We don't have the ability of fully flushing all caches (except perhaps with a device reset); but it seems we can flush at least the L2 cache - by allocating a large enough buffer and memset'ing it to 0. I hope it works.