On Wed, 2 May 2012 01:37:56 -0400, Devin Jeanpierre <[email protected]> wrote: > On Tue, May 1, 2012 at 10:02 AM, Andreas Kloeckner > <[email protected]> wrote: > > On Tue, 1 May 2012 02:58:33 -0400, Devin Jeanpierre > > <[email protected]> wrote: > >> On Tue, May 1, 2012 at 2:16 AM, Andreas Kloeckner > >> <[email protected]> wrote: > >> > Huh, strange. Does this happen with AMD CPU, too? What kernel/libc > >> > version? > >> > >> Kernel is 3.2.2-1.fc16.x86_64 , glibc is 2.14.90 > >> > >> I don't have any AMD machines available. In a bit over a week I'll > >> have an NVidia machine available, but that's it. > > > > AMD's CL implementation also works on CPUs, which might be worth > > trying. Also, while I don't think PyOpenCL has anything to do with your > > issue, have you tried using PyOpenCL git? > > Alright, it does halt on the AMD CPU, and gives the answer I wanted. > Woo! git's version of PyOpenCL doesn't make the Intel OCL > implementation halt though. > > With regards to Intel, what do I do next? And thanks for all your > help. :)
Honestly, this sounds like an Intel bug to me. My next step, if I was in your position, would be to try to reproduce the hang from C and then post to the Intel support forum. Here's some code to help you bring this up in C: http://forge.tiker.net/p/hw3/source/tree/master/ Andreas
pgpqpgGLmyCVl.pgp
Description: PGP signature
_______________________________________________ PyOpenCL mailing list [email protected] http://lists.tiker.net/listinfo/pyopencl
