Hello, users?
Currently, I'm trying to compile XOOPIC (2D plasma simulation program, MPI
parallel operation available) in conjunction with MPI.
I had no problem with XOOPIC compilation in single-machine operation mode;
however, when MPI mode is turned enabled, it generated about four or more
page
We recently got 33 new cluster nodes all of which have two onboard GigE
nics. We also got two powerconnect 2748 48 port switches which support
IEEE 802.3ad (link aggregation). I have configured the nodes to do
Ethernet bonding to aggregate the two nics in to one bonded device:
http://www.cyberci
There are 2 problems. First, it look like you're using LAM and not
Open MPI as there are some lam_ missing symbols. Second, please use
mpicxx to link your application as it will add all the missing
libraries.
george.
On May 24, 2007, at 10:38 AM, Jung, Soon-wook wrote:
Hello, users?
C
Hi,
For OpenMPI users struggling with cluster/grid setups in combination
with particular networks to use or avoid, it may be useful to know
that the documentation on http://www.open-mpi.org/faq/?category=tcp
regarding the oob module is currently misleading. The options that
actually implement th
To add to what George said -- it looks like you have multiple
different implementations of MPI installed on your machine (LAM/MPI,
MPICH, MPICH2, ...?). Ensure that you completely compile and run
your application with *one* implementation of MPI (they are not
binary compatible).
Keep in
On May 24, 2007, at 10:38 AM, Adams, Samuel D Contr AFRL/HEDR wrote:
We recently got 33 new cluster nodes all of which have two onboard
GigE
nics. We also got two powerconnect 2748 48 port switches which
support
IEEE 802.3ad (link aggregation). I have configured the nodes to do
Ethernet bo
You are absolutely right; the MCA parameter names between the oob and
btl tcp modules were slightly different (btl_tcp_if_[in|ex]clude and
oob_tcp_[in|ex]clude). And the FAQ is wrong -- I just updated it;
sorry about that, and thanks for pointing it out!
I literally just fixed this on the
Hi all,
Found a bug maybe:
I'm setting the following environment variables:
OMPI_MCA_btl="tcp,self"
OMPI_MCA_btl_tcp_if_include="eth2"
OMPI_MCA_btl_tcp_if_exclude="eth0"
However, when I run 'ompi_info --param btl tcp' I see (among other
things):
MCA btl: parameter "btl" (current value: )
MCA btl
It looks like this is a feature, not a bug. :-)
ompi_info is specifically clearing out the environment variables
corresponding to framework names (e.g., "btl"). If ompi_info didn't
do that, the OMPI core would only load the components that you have
specified in the environment and not sho
