Hi, When mpichg jobs are submitted to condor pool (of course through globus rsl file) with machine count = 1, jobs execute perfectly. However, when machine count > 1 is specified, job do not execute to the completion. I can see the jobs in the condor pool. Some of them are running and some of them are idle (eventhough machine are available for jobs to run, I could verify that using condor_status). This situation remains for a long time and jobs do not complete their execution. mpi program given for execution is simple helloworld program. Can you tell me what may be the problem? Thanks, Yadnyesh.
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}