Dear Amber users!
Using pmemd from latest Amber together with the newest version of mpirun
I've obtained very low performance in replica-exchange simulation using
multi-pmemd.mpi. Below you can find my launch bash script and the output
file from 1 replica.
export AMBERHOME=/opt/amber/amber14
pmemd="$AMBERHOME/bin/pmemd.
MPI"
sander="$AMBERHOME/bin/sander.MPI"
cd /globaltmp/novikov/amber/remd
mpirun -n 88 $pmemd -O -ng 22 -groupfile ./remd.groupfile2
| Average timings for last 100 steps:
| Elapsed(s) = 143.58 Per Step(ms) = 1435.78
| ns/day = 0.12 seconds/ns = 717888.57
|
| Average timings for all steps:
| Elapsed(s) = 143.58 Per Step(ms) = 1435.78
| ns/day = 0.12 seconds/ns = 717888.57
Here I performed simulation of the protein in vacuum (system of 7000 atoms)
using implicit solvent with the applied position restrains on the part of
my protein
Production REMD input file
&cntrl
irest=0, ntx=1,
nstlim=500, dt=0.002,
irest=0, ntt=3, gamma_ln=1.0,
temp0=300.00, ig=17461,
ntc=2, ntf=2, nscm=1000,
ntb=0, igb=5,
cut=999.0, rgbmax=999.0,
ntpr=100, ntwx=1000, ntwr=100000,
nmropt=1, ioutfm=1,
numexchg=1000,
restraint_wt=1000.0, restraintmask=':1-285,325-460',
/
&wt TYPE='END'
/
DISANG=chimera_chir.dat
Does the problem in the palatalization of this job or something wrong with
my input file?
TFH,
James
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jun 23 2014 - 05:00:02 PDT
