Lazzari Gianrocco <[email protected]> writes:

> Dear PYopenCL support,
>
>
>
> I just started use Pyopencl, by following this tutorial:
> https://www.google.ch/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0ahUKEwj0_dXOo7bKAhXBjQ8KHeg5AX8QFggdMAA&url=http%3A%2F%2Fwww.training.prace-ri.eu%2Fuploads%2Ftx_pracetmo%2FLinkSCEMM_pyOpenCL.pdf&usg=AFQjCNEHxox3fBDxCQ_VGAurAlH_cIKK7A&sig2=fyILhTUX8R5SaFCatUEMqg&cad=rja
>
> (this is also my first time with GPU parallel computing)
>
> I get some problems from the beginning:
>
> - cl.get_platforms() gives me:
>
> [<pyopencl.Platform 'Apple' at 0x7fff0000>]

That's actually not a problem--that indicates that everything is
working.

> - I tried also the 4  tests present in ~/pyopencl/test/
>
> with this (summarised) results:
>
> [cid:[email protected]][cid:[email protected]][cid:[email protected]]
>
> [cid:[email protected]]

Again, most of the tests pass, so you should be fine to get going. Some
of the more advanced tests hit snags with Apple's OpenCL
implementation. To get to the bottom of this, I would first run the
tests on CPU and GPU separately to better understand what's going on:

PYOPENCL_TEST=apple:cpu python -m pytest
PYOPENCL_TEST=apple:_____ python -m pytest

(you'll need to insert your GPU name for the blank)

If you're on OS X 10.10, at least the CPU tests should just pass. The
GPU tests may run into some (largely noncritical) issues.

If you're in the mood to help fix those, I'd welcome reports/patches.

Hope that helps,
Andreas

_______________________________________________
PyOpenCL mailing list
[email protected]
http://lists.tiker.net/listinfo/pyopencl

Reply via email to