AMBER Archive (2006)

Subject: Re: AMBER: AMBER 8: Problem with MPI_Finalize

From: David A. Case (
Date: Fri Jul 07 2006 - 16:29:10 CDT

On Fri, Jul 07, 2006, jrd wrote:
> I am running MD simulations of protein in water solvation using Amber 8 on
> SGI Altix 3700 supercomputer cluster running Novell's SUSE Linux Enterprise
> Server (SLES). The simulation box consisits of 11,000 atoms. The simulation
> ran for 28,000 timesteps, and then it was terminated with the following
> error message.
> -------
> MPI: MPI_COMM_WORLD rank 0 has terminated without calling MPI_Finalize()

This just says that the job stopped; you need to look at the output files
to see if there is any indication of what might have gone wrong. Altix
machines are usually pretty reliable, but it's possible that you had a
hardware problem -- you could check this if you want by re-running the job.


The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to