Ranger NEB error

From Henkelman Group

This page details the errors found when running an NEB job with climbing images on Ranger.

[edit] Set up

SYSTEM = 
IBRION=3
POTIM=0
IOPT=1
ISYM=0
NSW=1500
EDIFF=1E-05
EDIFFG=-0.001
PREC=normal
LWAVE=.FALSE.
LCHARG=.FALSE.
LREAL=AUTO
IMAGES=6
LCLIMB=.TRUE.
ENCUT=274

[edit] Results

Output written to ll_out
TACC: Setting memory limits for job 134376 to 1992294 KB
TACC: Dumping job script:
--------------------------------------------------------------------------------
#!/bin/tcsh
#
#$ -V
#$ -cwd
#$ -j y
#$ -N NEB_3
#$ -pe 16way 288
#$ -q normal
#$ -l h_rt=23:50:00
#$ -o ll_out

module unload mvapich2
module swap pgi intel
module load mvapich2


#setenv OMP_THREADS_NUM 1
ibrun /share/home/00405/graeme/bin/vasp

[some set up data and then printing ~60 iterations of expected data output]
Fatal error in MPI_Isend:
Internal MPI error!, error stack:
MPI_Isend(163): MPI_Isend(buf=0x2aaab7cc1d00, count=796, MPI_DOUBLE_PRECISION, d
est=42, tag=201, comm=0xc4000004, request=0x7fff1daea520) failed
(unknown)(): Internal MPI error!
rank 144 in job 1  i166-102.ranger.tacc.utexas.edu_35622   caused collective abo
rt of all ranks
  exit status of rank 144: killed by signal 9 
TACC: MPI job exited with code: 137
TACC: Shutting down parallel environment.
TACC: Shutdown complete. Exiting.
TACC: Cleaning up after job: 134376
TACC: Done.