Hi Mohsen,

On Thu, Sep 6, 2012 at 3:31 AM, mohsen jadidi <[email protected]> wrote:
>   File
> "/usr/local/lib/python2.7/dist-packages/scikits.cuda-0.042-py2.7.egg/scikits/cuda/linalg.py",
> line 323, in dot
>     c_gpu = gpuarray.empty((n, ldc), x_gpu.dtype)
>   File
> "/usr/local/lib/python2.7/dist-packages/pycuda-2011.2.2-py2.7-linux-x86_64.egg/pycuda/gpuarray.py",
> line 182, in __init__
>     self.gpudata = self.allocator(self.size * self.dtype.itemsize)
> pycuda._driver.MemoryError: cuMemAlloc failed: out of memory

The error looks pretty much self-explanatory. How many video memory do
you have? You can try to look into scikits/cuda/linalg.py and check
the value of "lds" to see what is the size of the array getting
allocated. Just estimating, if ldc == n, it is 6000 * 6000 * 8
(judging by the 'c' in the stack trace, it is a complex array)  ~ 250
Mb. And there may be several of them (again, judging by the name,
there may be "a_gpu" and "b_gpu"). If you have, say, a 512Mb-card, you
may be pushing the limit here.

_______________________________________________
PyCUDA mailing list
[email protected]
http://lists.tiker.net/listinfo/pycuda

Reply via email to