The latest CUDA release provides a function that allows to query the cudart
version:
cudaError_t cudaRuntimeGetVersion (int *runtimeVersion)
We should use it to detect and report mismatch errors.
Original issue reported on code.google.com by cast...@gmail.com on 24 Jul 2009 at 7:07
Original issue reported on code.google.com by
cast...@gmail.com
on 24 Jul 2009 at 7:07