[PyCUDA] Feature proposal: GPUArray views
nouiz at nouiz.org
Fri Sep 9 06:28:16 PDT 2011
This is in compyte for cuda. But there is not much more and there is
no conversion between compyte gpu nd array object and pycuda object
yet. There is no documentation for now.
So yes it is in the pipeline.
On Fri, Sep 9, 2011 at 3:44 AM, Bogdan Opanchuk <mantihor at gmail.com> wrote:
> Hi Andreas,
> On Fri, Sep 9, 2011 at 5:07 PM, Andreas Kloeckner
> <lists at informa.tiker.net> wrote:
>> PyCUDA already supports this--just do array[start:end]. PyOpenCL not
>> yet, but I'm meaning to add this at some point, unless compyte gets
>> there first.
> Thank you, that's great! The problem is that most of my arrays are
> multidimensional, and therefore cannot benefit from this.
> Moreover there seems to be now ravel() or reshape() equivalents in the
> interface, which could serve as a workaround (I checked the source
> now, just in case). So, I cannot really use it right now. Although the
> foundation is already there, so it may be easier for me to implement
> multidimensional slicing (unless you are planning to do it yourself in
> the near future).
> Best regards,
> PyCUDA mailing list
> PyCUDA at tiker.net
More information about the PyCUDA