[gmx-users] Restarts: Truncation of file *.xtc failed

[David de Sancho]

Dear Gromacs users
I am experiencing a problem with restarts in a REMD simulation with explicit
solvent in Gromacs 4.0.5. I am trying to continue a set of simulations that
finished after the maximum time in our cluster queue had been reached.
Therefore I just directly (i.e. without modifying my input file with
tpbconv) try continuing my run using

mpiexec "bindir"/mdrun_openmpi_intel -multi 32 -replex 5000 -v -s
tprfile.tpr -x xtcfile.xtc -o trrfile.trr  -c grofile.gro -e enefile.ene -g
logfile.log -cpi cptfile.cpt -cpo cptfile.cpt -append

This has worked before in previous continuations of this REMD simulation, so
in principle the pbs script I am using should be fine. However, now I find
two types of error messages:

(1) Reading checkpoint file cptfile{i}.cpt generated: ***
#PME-nodes mismatch,
current program: -1
 checkpoint file: 0
Gromacs binary or parallel settings not identical to previous run.
Continuation is exact, but is not guaranteed to be binary identical.

(2) Fatal error:
Truncation of file xtcfile${i}.xtc failed.

I have been browsing the lists and apparently the first problem is not very
important. However the second error, reached after loading the tpr and cpt
files, is fatal. I have thus checked what the problem was with my xtc files
by running gmxcheck on them and obtained messages like

# Atoms  6950
Precision 0.001 (nm)
Reading frame   90000 time 180000.016

Item        #frames Timestep (ps)
Step         90878    2
Time         90878    2
Lambda           0
Coords       90878    2
Velocities       0
Forces           0
Box          90878    2

Nothing here tells me much about what can be going wrong. Also, I have
looked at the last frame of the simulation using vmd in case there were
missing atoms in my protein but it definitely looks like there is nothing
wrong. I have tried with the cpt and prev.cpt files in case there was
something wrong with the cpt file but still I got the fatal error.
Any hints on what it is that I am doing wrong?
Thanks


David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100907/44f992ca/attachment.html>