On 11/20/2011 10:47 PM, Roy Stogner wrote: > > On Sun, 20 Nov 2011, David Knezevic wrote: > >> I was trying to debug ex30 in parallel, but it turns out on my >> current machine when I do, say, >> >> mpirun -np 2 ./ex4-dbg -n 10 -d 2 >> >> libMesh appears to run as 2 independent processes, i.e. n_elem() = >> n_local_elem() etc, rather than running in parallel. Have you ever >> encountered this before? I presume it's due to my mpich configuration >> or something (I just pulled it from synaptic... maybe I should >> compile from source). >> >> If you do know what might be causing that issue, it'd be interested >> to hear. > > Copying to libmesh-users in case others have seen this before with > different causes. > > The one time I've seen this before, it happened when I accidentally > had libMesh compiled against one MPI implementation and was running it > with mpirun or mpiexec with a different implementation.
Hi Roy, Yes, this was the issue, thanks. I thought I was using mpich2, but I also had openmpi installed (not sure if that was done by default by ubuntu or if I had installed it myself at some point in the distant past). Thanks, David ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
