Hi Lev, Yes, we are doing this in my project. We had the same issue as well, and found a workaround. Actually, we have been writing on the mailing list regarding this issue all weekend, and I think you will find some relevant code in my postings (gerald) showing the exact workaround we are using. If you can't find them I can repost, but I think they ought to be in the digest?
We are working to finish a python module of the entire CULA package, including device wrappers, and we are using pyCUDA for much of the device wrapper framework. I am hoping to have this ready soon; we are just tightening up a few issues and writing the testing routines now. Best of Luck, Garrett Wright Temple University ~ Mathematics
_______________________________________________ PyCUDA mailing list [email protected] http://lists.tiker.net/listinfo/pycuda
