Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] hwloc error in topology.c in OMPI 1.6.5
From: Brice Goglin (Brice.Goglin_at_[hidden])
Date: 2014-02-28 03:32:38

Le 28/02/2014 02:48, Ralph Castain a écrit :
> Remember, hwloc doesn't actually "sense" hardware - it just parses files in the /proc area. So if something is garbled in those files, hwloc will report errors. Doesn't mean anything is wrong with the hardware at all.

For the record, that's not really true:

hwloc looks at /sys (and a bit /proc files), but it also uses cpuid
instructions. 90% of the times, the former is better because the kernel
already took care of cleaning up the hardware mess and reporting
useful/correct info in /proc and /sys. Sometimes the kernel is too old
and it misses some hardware quirks (like L1i sharing on Gus' machine)
causing /sys files to be incompatible.

In the end, the vast majority of problems come from buggy BIOS, and
these cause both cpuid and kernel to report invalid info. Aside of
upgrading the BIOS, the only solution there is to replace the topology
with a correct XML one.