Hi Andreas,
On 06/22/11 07:02, Andreas Kloeckner wrote:
> I've just added an undocumented/preview feature to PyOpenCL git. You can
> now specify 'g_times_l=True' on a kernel invocation to get a more
> CUDA-like launch config specification, i.e. global size and global
> offset are multiplied by the local size.
>
> I'm debating whether to make this an official, documented feature (and
> thereby set it in stone) or leave it undocumented and thereby
> changeable.
I think this is a very convenient addition, in particular for people
moving from CUDA to OpenCL. The only problem I could see is that
somebody converting a PyOpenCL program to a C OpenCL program might
forget that he set g_times_l=True, but that's his problem.
All the best,
Jan
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDirig Dr. Karl Eugen Huthmacher
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Besuchen Sie uns auf unserem neuen Webauftritt unter www.fz-juelich.de
_______________________________________________
PyOpenCL mailing list
[email protected]
http://lists.tiker.net/listinfo/pyopencl
