Back to this painful issue, partly because  I found a workaround,
and partly because I would like to help.

 The initial post was :
where I reported about OMPI 1.4.1, but it was the same for 1.4.3.

 I spotted the culprit to be line #274 into btl_openib.c where it was required to replace
mca_btl_openib_component.qp_infos[qp].u.pp_qp.rd_rsv) * nprocs;
mca_btl_openib_component.qp_infos[qp].u.pp_qp.rd_rsv) * 32;
mostly because nprocs = 4096 or 8192 in our case, which was leading to a
very huge memlock resource requirement.

 Since I don't believe there is a relevant mca parameter to control this value accurately
(am I wrong ?), I would suggest to invent such switch.

 It occurs to work because the number of peers for a given node (apart for rank 0) is very low,
but it is definitely useful when all-to-all communication is not required on a big cluster.

 Could someone comment on this ?

 More info on request.

 Thanks,      Happy New Year to you all,       G.

Le 29/11/2010 16:58, Gilbert Grosdidier a écrit :
Bonjour John,

 Thanks for your feedback, but my investigations so far did not help:
the memlock limit on the compute nodes are actually set to unlimited.
This most probably means that even if the btl_openib hits some memory allocation
limit, the message I got is inaccurate because the memlock resource is indeed already unlimited.

 Then, the btl allocation mechanism seems to be stopped 
by the memlock resource being exhausted because the former is
attempting to create too many buffers, for example. I tried to explore this
kind of assumption by decreasing :
- btl_ofud_rd_num down to 32 or even 16
- btl_openib_cq_size down to 256 or even 64
but to no avail.

 So, I am asking for help about which other parameter could lead to (locked ?) memory exhaustion,
knowing that the current memlock wall shows up 
- when I run with 4096 or 8192 cores (for 2048, that's fine)
- there are 4GB of RAM available for each core
- each core is communicating with no more than 8 neighbours, and they
stay the same along the whole job life.

 Does this triggers some idea for anyone ?

 Thanks in advance,           Best,    Gilbert.

Le 20 nov. 10 à 19:27, John Hearns a écrit :

On 20 November 2010 16:31, Gilbert Grosdidier

Bonjour Gilbert.

I manage ICE clusters also.

Please could you have look at /etc/init.d/pbs on the compute blades?

Do you have something like:

   if [ "${PBS_START_MOM}" -gt 0 ] ; then
     if check_prog "mom" ; then
       echo "PBS mom already running."
       if [ -f /etc/sgi-release -o -f /etc/sgi-compute-node-release ] ; then
           MEMLOCKLIM=`ulimit -l`
           NOFILESLIM=`ulimit -n`
           STACKLIM=`ulimit -s`
           ulimit -l unlimited
           ulimit -n 16384
           ulimit -s unlimited
users mailing list

_______________________________________________ users mailing list

 Cordialement,   Gilbert.

  Gilbert Grosdidier   
  LAL / IN2P3 / CNRS                 Phone : +33 1 6446 8909
  Faculté des Sciences, Bat. 200     Fax   : +33 1 6446 8546
  B.P. 34, F-91898 Orsay Cedex (FRANCE)