NVE crash

From: pellegrini (pellegrini_at_ill.fr)
Date: Wed Apr 28 2010 - 10:30:09 CDT

Dear NAMD users,

I was running some performance tests for a simple system and I fell on
strange results I can not explain.

I created a waterbox of 50 A^3 using the solvation box command in VMD.
It populated it with ~11000 atoms.

I performed two tests:

    Test1: Minimization of 100 steps followed by NVE 10 ps at 300K
    Test2: Directly starts with NVE 10ps at 300K

For Test1, everything ran OK but Test2 immediatly crashed because of an
undefined Pressure and GPressure.
Here is the difference obtained for the two tests:

Test1:

Info: Entering startup phase 8 with 10588 kB of memory in use.
Info: Finished startup with 10588 kB of memory in use.
TCL: Minimizing for 100 steps
PRESSURE: 0 450956 14673.7 22989.7 14673.7 938902 -205247 22989.7
-205247 139014
GPRESSURE: 0 512552 7544.79 129989 12092.8 972608 -174860 32009.9
-211810 135335
ETITLE: TS BOND ANGLE DIHED
IMPRP ELECT VDW BOUNDARY
MISC KINETIC TOTAL TEMP
TOTAL2 TOTAL3 TEMPAVG PRESSURE GPRESSURE
       VOLUME PRESSAVG GPRESSAVG

ENERGY: 0 3645.1246 849.7672 0.0000
0.0000 -40209.2398 234284.2838 0.0000 0.
0000 0.0000 198569.9358 0.0000 198569.9358
198569.9358 0.0000 509624.3505 540164.9075
  125000.0000 509624.3505 540164.9075

Test2:

Info: Finished startup with 10896 kB of memory in use.
TCL: Running for 10000 steps
PRESSURE: 0 nan nan nan nan nan nan nan nan nan
GPRESSURE: 0 nan nan nan nan nan nan nan nan nan
ETITLE: TS BOND ANGLE DIHED
IMPRP ELECT VDW BOUNDARY
MISC KINETIC TOTAL TEMP
TOTAL2 TOTAL3 TEMPAVG PRESSURE GPRESSURE
       VOLUME PRESSAVG GPRESSAVG

ENERGY: 0 0.0000 0.0000 0.0000
0.0000 -37531.6643 235723.3443 0.0000 0.
0000 99999999.9999 99999999.9999 99999999.9999 99999999.9999
99999999.9999 99999999.9999 99999999.9999 99999999.9999
  125000.0000 99999999.9999 99999999.9999

Would you have any idea why I got this difference ? Normally I would
expect the first step that consists in a single energy calculation to be
identical for both test. No ?

Attached is the configuration script used for both test (for Test1 the
final 'minimize' is just commented out).

thank you very much

Eric

source parameters.txt

# input
coordinates ${pdb}
structure ${psf}
parameters ${parameters}
paratypecharmm on

# output
outputname output
dcdfile output.dcd
xstFile output.xst
dcdfreq 100
xstFreq 100

binaryoutput no
binaryrestart no
outputEnergies 100
outputPressure 100

# Basic dynamics
exclude scaled1-4
1-4scaling 1
COMmotion no
dielectric 1.0
rigidBonds all

# Simulation space partitioning
switching on
switchdist 10
cutoff 12
pairlistdist 13.5

# Multiple timestepping
firsttimestep 0
timestep 1
stepspercycle 20
nonbondedFreq 2
fullElectFrequency 4

# Temperature control
temperature 300

# PBC
cellBasisVector1 50.0 0.0 0.0
cellBasisVector2 0.0 50.0 0.0
cellBasisVector3 0.0 0.0 50.0
cellOrigin 25.0 25.0 25.0
wrapAll on
dcdUnitCell yes

PME yes
PMEGridSpacing 1.0

# Scripting
#minimize 100
run 10000

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:55:43 CST