Open MPI logo

Hardware Locality Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Hardware Locality Development mailing list

Subject: Re: [hwloc-devel] [hwloc-svn] svn:hwloc r3157
From: Samuel Thibault (samuel.thibault_at_[hidden])
Date: 2011-02-04 06:57:55


Brice Goglin, le Fri 04 Feb 2011 12:03:50 +0100, a écrit :
> > Mmm, I then wonder whether we should really provide a flag, then, if
> > it's a matter of calling restrict() afterwards. What would be more
> > generic and useful is a topology configuration option which restricts
> > detection to a given cpuset: the core filters out objects outside from
> > that cpuset, and OS backends can also avoid trying to detect them,
> > making detection faster.
>
> Right, we can optimize these things later. The above is just an easy implem.

Sure, but the very fact that it can be easily obtained another way makes
me wonder whether it is worth adding a flag when there is another way.
(I'm not saying the flag shouldn't be added, but just whether it's worth
it).

Samuel