On 14.02.2013, at 18:36, Brice Goglin <Brice.Goglin_at_[hidden]> wrote:
> Le 14/02/2013 14:00, Stefan Eilemann a Ã©crit :
>> On 12. Feb 2013, at 23:37, Brice Goglin <Brice.Goglin_at_[hidden]> wrote:
>>> Stefan (or anybody else interested in hwloc GPU support),
>>> Did you have any chance to look at this?
>> I can confirm that it works for us now. I did implement an important performance optimization, please look at the usual bug for the patch: https://github.com/BlueBrain/hwloc/issues/1
> Can you explain a bit more what you did? I understand that you use
> ScreenCount() to avoid useless tests.
I use the same display connection for all screens, doing 10 XOpenDisplay instead of a 100.
> But you're only testing the NV-CONTROL extension on :X, not on each
> :X.Y. Is it guaranteed that :0.0, :0.1 ... support the same list of
Yes, the extensions are per server, not per screen.
> Isn't it possible to have a single X server with a NVIDIA
> board driving :0.0 and a non-NVIDIA board driving :0.1 ?
Good question. I haven't seen this, so I've got no clue if one would get the union or intersection of extensions. Maybe it's not even possible?
> By the way, please try to not reindent the code next time, it makes the
> patch much harder to read :/ (your editor may be doing it in your back)
I didn't. I just untabified the code to make the indentations line up.