Thanks to you both!

I figured it was something like that. That was exactly what I needed.

Gratefully yours,

Hugh

On Sun, Feb 12, 2012 at 4:32 PM, Bogdan Opanchuk <[email protected]> wrote:
> Hello Hugh,
>
> On Mon, Feb 13, 2012 at 9:41 AM, Hugh Owens <[email protected]> wrote:
>> aa = np.ones(512)
>
> Here aa has dtype=np.float64 (default). You have to explicitly convert
> it to an array of floats:
>
>> aa = np.ones(512).astype(np.float32)
>
> Best regards,
> Bogdan

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

Reply via email to