Silly me ! I've just seen that Andrej reported exactly the same bug last month. 
I checked his .output file and it seems he got the same hardware than me. I see 
now why you said "yet another buggy AMD platform" !

Sorry guys.


List-Post: hwloc-users@lists.open-mpi.org
Date: Fri, 5 Sep 2014 13:46:25 +0200
From: brice.gog...@inria.fr
To: hwloc-us...@open-mpi.org
Subject: Re: [hwloc-users] hwloc error with "node interleaving" disabled


  
    
  
  
    Hello

      

      You sent the test.output file instead of test.tar.bz2 so I can't
      check for sure. Anyway I guess this is yet another buggy AMD
      platform with magny-cours/interlagos/abu-dahbi Opterons (61xx,
      62xx or 63xx). 

      

      Sometimes upgrading the BIOS/kernel helps. Sometimes not.

      

      Some L3 caches will be missing in the hwloc topology because of
      this bug, it's likely not important for the vast majority of HPC
      libraries.

      

      You may hide the warning by setting HWLOC_HIDE_ERRORS=1 in your
      environment.

      

      Brice

      

      

      

      

      Le 05/09/2014 12:06, Jean-Pierre Adam a écrit :

    
    
      
      Hello hwloc experts

        

        I encounter this bug when I'm using mpirun or hwloc directly :

        

****************************************************************************

        * hwloc has encountered what looks like an error from the
        operating system.

        *

        * L3 (cpuset 0x000003f0) intersects with NUMANode (P#0 cpuset
        0x0000003f) without inclusion!

        * Error occurred in topology.c line 940

        *

        * Please report this error message to the hwloc user's mailing
        list,

        * along with the output from the hwloc-gather-topology script.

****************************************************************************

        

        The output of hwloc-gather-topology is attached. The OS is
        Centos 7.

        

        The tool launched with mpirun runs as expected, still the
        message is a bit worrying...

        

        I was able to avoid this message by enabling "node interleaving"
        in the bios (basically disables NUMA). In my case, I got a 5%
        performance loss with that setting. It could be acceptable, but
        I would like to understant what is going on.

        

        So is my motherboard / BIOS / OS buggy ?

        

        Best regards

      
      

      
      

      _______________________________________________
hwloc-users mailing list
hwloc-us...@open-mpi.org
Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users
Link to this post: 
http://www.open-mpi.org/community/lists/hwloc-users/2014/09/1070.php
    
    

  


_______________________________________________
hwloc-users mailing list
hwloc-us...@open-mpi.org
Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users
Link to this post: 
http://www.open-mpi.org/community/lists/hwloc-users/2014/09/1072.php            
                               

Reply via email to