Thanks Andreas, I've checked with a longer running kernel, and there is still no overlap. Now I'm going to re-install pycuda and cuda4.0, enable api tracing and go from there. Thanks!
Jesse On Mon, Apr 4, 2011 at 9:12 PM, Andreas Kloeckner <li...@informa.tiker.net>wrote: > On Mon, 4 Apr 2011 16:35:22 -0700, Jesse Lu <jess...@stanford.edu> wrote: > > Hi everyone, > > > > I have been trying to create a kernel concurrency example for pycuda but > > have been unsuccessful. Can anyone point out what I'm doing wrong? Thanks > in > > advance! > > Two suggestions: > > - Perhaps try using longer-running kernels. (a simple for loop in the > kernel might do the trick) > > - Enable tracing in PyCUDA to see if any allocations or deletions happen > while the kernel loop is going. (Those are often implicit sync points > in CUDA.) It doesn't look like it, but it can't hurt to check. > > Andreas >
_______________________________________________ PyCUDA mailing list PyCUDA@tiker.net http://lists.tiker.net/listinfo/pycuda