Received from reckoner on Thu, Dec 16, 2010 at 10:12:06PM EST: > On my Linux box, I have finally figured out that when I do > > import pycuda.autoinit > import scikits.cuda.linalg as culinalg > culinalg.init() > > I get a > > scikits.cuda.cublas.cublasNotInitialized > > exception. > > However, if I do > > import scikits.cuda.linalg as culinalg > culinalg.init() > import pycuda.autoinit > > Note that the order of the statements has changed, but at least I no > longer get the "cublasNotInitialized" error anymore. By the way, > I've been trying to get the
I can't reproduce this with CUDA 3.2 on 64-bit Linux. What happens if you try import pycuda.autoinit import scikits.cuda.cublas as cublas cublas.cublasInit() > svd_demo.py > > from scikits.cuda-0.03/demos working. Additionally, > > pycuda.VERSION=94 > > and > > GeForce 9800M GTS > Linux 2.6.31-22-server #69-Ubuntu SMP x86_64 GNU/Linux > > with the latest CUDA SDK. You mean 3.2? > Any help appreciated! You can submit bug reports regarding scikits.cuda at http://github.com/lebedov/scikits.cuda L.G. _______________________________________________ PyCUDA mailing list [email protected] http://lists.tiker.net/listinfo/pycuda
