Open MPI logo

Hardware Locality Development Mailing List Archives

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

Subject: Re: [hwloc-devel] hwloc-1.3.1 assertion failures on Linux/POWER7
From: Samuel Thibault (samuel.thibault_at_[hidden])
Date: 2012-02-02 05:54:24

Paul H. Hargrove, le Thu 02 Feb 2012 01:43:45 +0100, a écrit :
> On 2/1/2012 4:14 PM, Christopher Samuel wrote:
> >On 02/02/12 10:38, Paul H. Hargrove wrote:
> >
> >>> I am not sure if one should fix this by:
> >>> a) Document the need for CFLAGS=-qhalt=e
> >>> b) Force "-qhalt=e" at configure time when CC=xlc
> >>> c) Find some other way to fix the configure probe
> >>> > My vote is for "(b)"
> >Mine too, either that or:
> >
> >d) Exit saying the user has to set it, or use a different compiler.
> I think I'd like to change my vote to (d) IFF we can determine the flag is
> needed but not set.
> Something like (untested):

I would actually say both: pass -qhalt=e to xlc, and also check in all
cases that the compiler does check for argument count. Perhaps also
add a configure option to force either setschedaffinity way, in case a
compiler does not have an option to detect argument count issue.