Implicit MD is slower than explcit MD!

From: Krishna Vkm (krishnatifr_at_gmail.com)
Date: Mon Nov 23 2020 - 05:18:51 CST

Dear all
I have simulated a peptide in the explicit solvent in NAMD 2.13 and it took
around 2.9 hr to complete 10 ns. I needed a cheaper simulation for the same
system, so I went for implicit MD. However, in implicit MD (using GBIS) of
the peptide, the speed is even slower where it took 3.7 hr to complete
10ns. During the explicit and implicit run, the cluster condition is the
same (same number of nodes and processor and also no other jobs were
running) so it appears that there could be something in the configuration
file or there is something else. I am giving below the configuration file
for implicit and explicit MD.
I will be really thankful for any help

Regards
Krishna

*####Configuration file for explicit MD#####*

coordinates npt-prot-free.coor
structure step3_charmm2namd.psf
extendedSystem npt-prot-free.xsc
velocities npt-prot-free.vel
# forcefield
parameters toppar/par_all36m_prot.prm
parameters toppar/toppar_water_ions.str
paraTypeXplor off
paraTypeCharmm on
exclude scaled1-4
1-4scaling 1.0
# output files
outputname nvt-prod-50ns-1
restartname nvt-prod-50ns-1-restart
restartfreq 1000000
restartsave yes
binaryoutput no
DCDfile nvt-prod-50ns-1.dcd
DCDfreq 1000
outputEnergies 100
outputMomenta 1000
outputTiming 1000
#approximations
cutoff 12.0
switching on
switchdist 10.0
pairlistdist 14.0
splitPatch hydrogen
hgroupcutoff 2.5
PME yes
PMEGridSpacing 1
#constraints
rigidBonds all # Reqd for a 2fs timestep.
rigidTolerance 0.00001
rigidIterations 100
#integrator
numsteps 500000
timestep 2.0 # 2 fs timestep
stepspercycle 10
# simulation parameters
COMmotion no
dielectric 1.0
fullElectFrequency 1
nonbondedFreq 1
longSplitting c1
#temp control
langevin on
langevinTemp 300.0
langevinDamping 5
langevinHydrogen on
#script
run 5000000

*####Configuration file for implicit MD#####*

structure implicit.psf
coordinates prot-free.coor
velocities prot-free.vel

# forcefield
parameters toppar/par_all36m_prot.prm
paraTypeXplor off
paraTypeCharmm on
exclude scaled1-4
1-4scaling 1.0
# output files
outputname prod-50ns-1
#restartname prod-50ns-1
restartfreq 5000000
restartsave yes
binaryoutput no
#DCDfile prod-50ns-1.dcd
DCDfreq 1000
outputEnergies 10
outputMomenta 1000
outputTiming 1000
# GBIS PARAMETERS
gbis on
ionConcentration 0
sasa on
surfaceTension 0.005
#approximations
cutoff 16.0
switching on
switchdist 15.0
pairlistdist 18.0
splitPatch hydrogen
hgroupcutoff 2.5
#constraints
rigidBonds all
rigidTolerance 0.00001
rigidIterations 100
#integrator
numsteps 500000
timestep 2.0
stepspercycle 10
# simulation parameters
#temperature 300.0
COMmotion no
dielectric 1.0
fullElectFrequency 1
nonbondedFreq 1
longSplitting c1
#temp control
langevin on
langevinTemp 300.0
langevinDamping 5
langevinHydrogen on
run 5000000

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:14 CST