using fftconvolve2d() #1
Description
Hello,
I'm trying to use python-cuda with CUDA 2.3. Unfortunately, I don't have root privileges so I'm stuck with 2.3. I receive the following error at run-time.
Any ideas?
Thanks!
Brian
Compiling kernel using the following options:
nvcc --shared -Xcompiler='-fPIC' -o /tmp/tmp5Z849H/kernel.so kernel.cu
Traceback (most recent call last):
File "doit.py", line 13, in
cuda.sugar.fft.fftconvolve2d(data, kernel, True)
File "/home/ascdsdev/brefsdal/lib64/python2.5/site-packages/cuda/sugar/fft/fftconvolve.py", line 155, in fftconvolve2d
padKernel = fftconvolve2d.padKernel(kernelBlockGrid, threadBlock)
File "/home/ascdsdev/brefsdal/lib64/python2.5/site-packages/cuda/sugar/kernel/kernelfactoryrt.py", line 28, in getattr
raise AttributeError("could not find kernel named %r in %r" % (name, self.dll))
AttributeError: could not find kernel named 'padKernel' in <CDLL '/tmp/python-cuda-compiler-cache-v1-uid501/5bf451004457afb5d7017793e911c48d.so', handle 2cd2da0 at 2c23490>