[gmx-users] Extending a run for different number of nodes
DimitryASuplatov
genesup at gmail.com
Mon Sep 15 10:09:15 CEST 2008
Hello,
I am sorry to ask this again but I guess that nobody was reading mail on
sunday.
I have completed a 10ns run. Now I want to extend it to 20ns. The
problem is that the initial tpr was compiled for 8 processors and now
I have only 4 cores.
What I tried to do is used tpbconv to make a 10ns extended tpr, then
used an exactly the same mdp file to grompp it all together with -np 4
to a new tpr file.
The problem is that at the beginning of the trajectory the structure
is undergoing a structural change with 0.12 nm from the initial
structure. That means that the structure is actually relaxing and that
is not something I want to see.
I have used this in my mdp
;------------ VELOCITY INITIATION -----------------------
gen-vel = no
gen-temp = 300
That means that no velocities were generated and I expect to see no
relaxation but the continuation of my run.
What am I doing wrong? How to extend a run without relaxation.
Thank you. I appreciate any help.
SDA
My mdp file is the following.
;-------------PREPROCESSING OPTIONS------------
title = PR "1KLC FREE"; my title
cpp = cpp; my C prepocessor (whereIsIt)
include = ; dir to include in my topology file
define = ;
;-------------RUN CONTROL---------------------
integrator = md;
;--------------OUTPUT CONTROL------------------
;TRR outpur
nstxout = 10000
nstvout = 10000
nstfout = 0
;LOG and EDR output
nstlog = 10000
nstenergy = 10000
;XTC output
nstxtcout = 1000
xtc-precision = 1000
;-------------START TIME & TIMESTEP-----------
tinit = ; [ps] starting time for your run (only makes
sense for md, sd, bd integator)
dt = 0.005 ; [ps] time step for integration (--"--)
nsteps = 1600000 ; (0) maximum number of steps to
integrate
comm-mode = ; Linear(Remove center of mass
translation)/Annular(Remove center of mass
;translation and rotation)/No(no restriction)
nstcomm = 1
;-------------NEIGHBOR SEACHING---------------
nstlist = 5; frequency o update
neighbor list
ns_type = grid; make a gid in
the box and check only
;neighboring grid cell when constructing a new neighbor list
rlist = 1.0; cut-off distance
for the short-range neighbor list
;-------------ELECTROSTATICS------------------
coulombtype = Reaction-field ; reaction field with
coulomb cut-off rcoulomb
; where rcoulomb>rlist. The dielectric constant beyond the cut-off is
epsilon_r
rcoulomb = 1.5 ;distance for the
coulomb cut-off
epsilon_rf = 80;dielectric constant
epsilon_r = 1;
vdwtype = cut-off; twin-range
cut-off`s with neighbor list cut-off rlist
;and VdW cut-off rvdw, where rvdw>rlist
rvdw = 1.0 ; distance for for
the LJ or Buckingam cut-off
;-------------ENERGY MINIMIZATION-------------
emtol = 200 (kJ/mol*nm); minimisation is converged when
the maximum force is smaller then this value
emstep = 0.01; initial step size
;-------------TEMTERAURE CONTROL--------------
tcoupl = berendsen
tc-grps = System
tau-t = 0.1
ref-t = 300
;--------- PRESSURE CONTROL---------------------
Pcoupl = no
Pcoupltype = Isotropic
tau-p = 1
ref_p = 1
compressibility = 4.5E-5
;------------ VELOCITY INITIATION -----------------------
gen-vel = no
gen-temp = 300
;------------- BONDS -----------------------
constraints = all-bonds
constraint-algorithm = Lincs
unconstrained-start = yes
Shake-SOR = no
shake-tol = 0.0001
lincs-order = 8
lincs-warnangle = 30
lincs-iter = 8
morse = no
More information about the gromacs.org_gmx-users
mailing list