I was wondering if there was a rough estimate of when / what would
trigger 3d grids being supported in pycuda (is it only formal removal
of the stateful interface in cuda that would cause this to happen?).
We are running into issues where the real solution is to have a Z-grid
1, and I'm trying to determine if implementing a
workaround for only
having 2d grids is worthwhile. Thanks for any info!
Cheers,
Eli
PS - Apologies for the thread necromancy.
On Mon, Dec 12, 2011 at 8:17 PM, Andreas Kloeckner
<lists(a)informa.tiker.net> wrote:
On Mon, 12 Dec 2011 21:37:38 -0500, Yifei Li
<yifli82(a)gmail.com> wrote:
Hi all,
It is said in CUDA 4.0 programming guide that "Blocks are organized into a
one-dimensional, two-dimensional, or three-dimensional
grid of thread blocks". Does PyCuda currently support 3-dimensional
grid?
Nope, this will happen once PyCUDA switches over to stateless launches
internally, as described in my previous email. CUDA's stateful launch
interface is 2D-grid-only, and PyCUDA uses that internally for greater
backwards compatibility.
Andreas
_______________________________________________
PyCUDA mailing list
PyCUDA(a)tiker.net
http://lists.tiker.net/listinfo/pycuda