I am currently running a case on an HPC system. The case is successful when all 64 processors are assigned to a single node. However, it fails when distributing the processors across two nodes. In this scenario, the case does not run with any error output, nor does it terminate automatically. Upon examining the log file, I noticed that there is no "done initmpi" message, leading me to suspect an issue with the MPI initiation process.
I am currently running a case on an HPC system. The case is successful when all 64 processors are assigned to a single node. However, it fails when distributing the processors across two nodes. In this scenario, the case does not run with any error output, nor does it terminate automatically. Upon examining the log file, I noticed that there is no "done initmpi" message, leading me to suspect an issue with the MPI initiation process.