dear members of the amber community,
I have encountered the following problem during a production run (npt).
I have started from an peptide in a (hopefully) carefully equilibrated
methanol box. After the last minimization I have gradually warmed up the
system. Temp and pressure are behaving "normally". After some time the
output looks the follows:
-- NSTEP = 18900 TIME(PS) = 42.200 TEMP(K) = 320.26 PRESS =
-9.23
Etot = 6259.0534 EKtot = 4985.0872 EPtot =
1273.9661
BOND = 1177.9108 ANGLE = 1670.5046 DIHED =
345.1988
1-4 NB = 36.2095 1-4 EEL = 3869.7420 VDWAALS =
-1695.1904
EELEC = -4130.4091 EHBOND = 0.0000 CONSTRAINT =
0.0000
EKCMT = 0.8179 VIRIAL = 14.5330 VOLUME =
68847.9731
Density =
0.6791
Ewald error estimate: 0.8250E-04
------------------------------------------------------------------------------
check COM velocity, temp: 0.000000 0.00(Removed)
check COM velocity, temp: 0.000000 0.00(Removed)
check COM velocity, temp: 0.000000 0.00(Removed)
check COM velocity, temp: nan nan(Removed)
check COM velocity, temp: nan nan(Removed)
check COM velocity, temp: nan nan(Removed)
NSTEP = 19000 TIME(PS) = 42.300 TEMP(K) = nan PRESS =
nan
Etot = nan EKtot = nan EPtot =
nan
BOND = nan ANGLE = 388307.1723 DIHED =
0.0000
1-4 NB = nan 1-4 EEL = nan VDWAALS =
-nan
EELEC = nan EHBOND = 0.0000 CONSTRAINT =
0.0000
Ewald error estimate: 0.0000E+00
I see no indications that something may have gone wrong in the prior
steps. I can circumvent the problem if I reduce the timestep to 0.2 fs,
but this is unrealistic for calculating longer trajectories. More
astonishingly, the calculation goes on without crash (what is calculated
afterwards ?).
Is the system non-equilibrated? Is is a numerical problem (wrong
compilation)? Where do I have to look for the problem?
This is the AMBER input file:
&cntrl
imin=0,ntx=7,irest=1,nmropt=0,ntwprt=169,
ntpr=100,ntwx=1000,ntwv=1000,ntwe=1000,
ntf=1,ntb=2,dielc=1.0,
ibelly=0,ntr=0,
nstlim=500000,dt=0.001,
temp0=320.0,
ntt=1,tautp=1.0,ndfmin=6,ntcm=1,nscm=15,
ntp=1,taup=0.2,npscal=1,
ntc=1,
comp=144.5,
&end
(The calc. is done on a linux version comp. with the pg-f77 compiler ,
gnu gives similar results)
With kindest regards,
Oliver
___
Dr. Oliver Zerbe
Department of Applied Biosciences
ETH-Zuerich
Winterthurerstr. 190
CH-8057 Zuerich
Tel.: 0041-1-635 60 81/ 635 6080
Fax: 0041-1-635 68 84
mailto:oliver.zerbe_at_pharma.anbi.ethz.ch
WWW: http://www.pharma.ethz.ch/people/oliver.zerbe/
Received on Wed Oct 03 2001 - 09:45:10 PDT
