> On Mar 18, 2019, at 9:14 AM, Brice Goglin <brice.gog...@inria.fr> wrote:
> I am planning to fix this in 2.1 (to be released before summer). I'll
> backport the trivial pieces to 2.0.x too.
> Do you care about a specific value passed to -Wstack-usage=X? or do you
> just want to avoid dynamic/unbounded allocs on the stack?

Thanks, Brice.  The specific value is not so important as long as it's not 
crazy high.  FYI, we use 256K in mpich, but that was somewhat arbitrarily 

  -- Pavan

hwloc-users mailing list

Reply via email to