That's extremely strange. There's a bug there to fix (48 cannot be correct in both places it is used). Albert, can you please upload your .tpr to a new issue at http://redmine.gromacs.org? Please also add what value of OMP_NUM_THREADS is set (perhaps implicitly by the script or your mpirun or mpirun setup).
Mark On Tue, Jan 28, 2014 at 11:12 AM, Albert <mailmd2...@gmail.com> wrote: > Hello: > > I am submitting a gromacs job in a cluster with command: > > > module load module load gromacs/4.6.5-intel13.1 > export MDRUN=mdrun_mpi > #steepest MINI > mpirun -np 1 grompp_mpi -f em.mdp -c ion-em.gro -p topol.top -o em.tpr -n > mpirun -np 64 mdrun_mpi -s em.tpr -c em.gro -v -g em.log &>em.info > > > but it failed with messages: > > > ------------------------------------------------------------ > ---------------------------- > Reading file em.tpr, VERSION 4.6.5 (single precision) > > Will use 48 particle-particle and 16 PME only nodes > This is a guess, check the performance at the end of the log file > Using 64 MPI processes > Using 12 OpenMP threads per MPI process > > ------------------------------------------------------- > Program mdrun_mpi, VERSION 4.6.5 > Source code file: /icm/home/magd/gromacs-4.6.5/src/mdlib/nbnxn_search.c, > line: 2523 > > Fatal error: > 48 OpenMP threads were requested. Since the non-bonded force buffer > reduction is prohibitively slow with more than > 32 threads, we do not allow this. Use 32 or less OpenMP threads. > For more information and tips for troubleshooting, please check the GROMACS > website at http://www.gromacs.org/Documentation/Errors > ------------------------------------------------------- > > "Everybody Lie Down On the Floor and Keep Calm" (KLF) > > Error on node 1, will try to stop all the nodes > Halting parallel program mdrun_mpi on CPU 1 out of 64 > > gcq#78: "Everybody Lie Down On the Floor and Keep Calm" (KLF) > > -------------------------------------------------------------------------- > MPI_ABORT was invoked on rank 1 in communicator MPI_COMM_WORLD > with errorcode -1. > > NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. > You may or may not see output from other processes, depending on > exactly when Open MPI kills them. > -------------------------------------------------------------------------- > -------------------------------------------------------------------------- > > -- > Gromacs Users mailing list > > * Please search the archive at http://www.gromacs.org/ > Support/Mailing_Lists/GMX-Users_List before posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. > -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
