Hi, I have recently had many problems with jobs that say they exceed memory and terminate, yet I am allocating 128gb of memory for a system which theoretically should need less than this. It just randomly terminates even if less memory if being used and I was wondering what the error was.
Here is the input file:
memory 128 Gb
molecule complex {
0 2
C -2.83891 -0.84747 -0.20778
N -2.06959 -0.27604 0.77905
C -0.85937 -0.82512 0.74757
N -0.82980 -1.74677 -0.21226
C -2.05910 -1.77487 -0.83179
C -2.45490 0.86620 1.59959
C 0.35903 -2.49473 -0.61265
F -1.01767 1.46174 -1.11980
B 0.28985 1.63764 -0.63648
F 0.23738 1.82826 0.77419
F 0.89453 2.73957 -1.24020
F 1.05039 0.46734 -0.88673
H -0.02604 -0.54755 1.38339
H -2.27287 -2.45330 -1.64959
H -3.86751 -0.55086 -0.37733
H -3.45079 0.68580 2.01748
H -1.72185 0.98138 2.40249
H -2.44673 1.75972 0.96603
H 0.98379 -1.85263 -1.24323
H 0.91655 -2.78676 0.28358
H 0.03922 -3.39046 -1.15377
C 2.81822 0.12352 1.52490
C 3.26768 -0.78277 0.56145
H 3.28632 0.11977 2.50813
H 2.01109 0.81253 1.27433
N 3.64537 -1.55240 -0.23308
}
set globals {
basis aug-cc-pVQZ
scf_type df
freeze_core True
basis_guess cc-pVDZ
guess sad
S_Orthogonalization=canonical
reference rohf
}
set cachelevel 0
set print 2
energy(‘dfmp2’)
Here is the job file:
#!/bin/bash
#PBS -P k96
#PBS -q normal
#PBS -l mem=128GB
#PBS -l ncpus=32
#PBS -l jobfs=500GB
#PBS -l walltime=4:00:00
#PBS -j oe
#PBS -l wd
module load psicode/4.0b5.1
export OMP_NUM_THREADS=32
export MKL_NUM_THREADS=32
psi4 ch2cnmimbf4conf4conf1avqzsummer.inp ch2cnmimbf4conf4conf1avqzsummer.out
Here is the .o file:
/jobfs/local/pbs/mom_priv/jobs/5883868.r-man2.SC: line 14: 19944 Killed psi4 ch2cnmimbf4conf4conf1avqzsummer.inp ch2cnmimbf4conf4c$
Process psi4 (PID 19944) was killed by kernel for exceeding memory limits. VM: 80570024kB, RSS: 61485776kB, File RSS: 1180kB
======================================================================================
Resource Usage on 2016-04-28 05:39:41:
Job Id: 5883868.r-man2
Project: k96
Exit Status: 137 (Linux Signal 9 SIGKILL Kill, unblockable)
Service Units: 18.07
NCPUs Requested: 32 NCPUs Used: 32
CPU Time Used: 07:56:48
Memory Requested: 128.0GB Memory Used: 54.95GB
Walltime requested: 04:00:00 Walltime Used: 00:33:53
JobFS requested: 500.0GB JobFS used: 581.68MB
This is not system size specific and some large systems have run while smaller systems have not, in general with the 4.0b5.1 psi4 I can get not augmented basis sets to work while augmented ones don’t once larger than double zeta. Does anybody have any suggestions?