[gmx-users] Fwd: gmx velacc does not calculate on "too often saved" files?

[Jones de Andrade]

Dear all,

I'm resending this message, since I got no answer.

I'm issuing the following command line:

gmx velacc -f /misc/lustre1/jdandrade/1000.BMIm.AlCl4.md.005.trr -o
./analises/1000.BMIm.AlCl4.md.005.velacc.vacfANI.xvg -mol -s
1000.BMIm.AlCl4.md.005.tpr -normalize

My system has 30 thousand atoms. The .trr file is saved with coordinates,
velocities and forces in it.

The command provided gives me a null VACF in case of nstxout, nstvout and
nstfout are all set to 1, 2, 5 or 10 in the .mdp file. It does give me a
non-null result when these parameters are 100 (but this is useless in my
case).

It happens consistently in simulations with nsteps set to 100000, 250000
and 500000.

Has anybody here ever come across such issue?

Thanks a lot in advance,

Jones

---------- Forwarded message ----------
From: Jones de Andrade <johannesrs at gmail.com>
Date: Thu, Oct 20, 2016 at 1:58 PM
Subject: gmx velacc does not calculate on "too often saved" files?
To: Discussion list for GROMACS users <gmx-users at gromacs.org>
Cc: gromacs.org_gmx-users at maillist.sys.kth.se


Dear Gromacs users,

I'm having some issues with gmx velacc now.

My systems have two molecules mixed, 1000 of each, one with 5 atoms and
other with 25: 30,000 atoms in total.

My issued command line reads:

gmx velacc -f /misc/lustre1/jdandrade/1000.BMIm.AlCl4.md.005.trr -o
./analises/1000.BMIm.AlCl4.md.005.velacc.vacfANI.xvg -mol -s
1000.BMIm.AlCl4.md.005.tpr -normalize

I'm testing gmx velacc with different number of time steps and recording
frequency on the .trr file (for other reasons and analysis, it's being
writen with positions, velocities and forces).

What I'm observing is that gmx velacc outputs me fully zeroed .xvg files
for either molecule if I use it against a .trr file that is recorded every
time step, every two time steps, every five time steps or every ten time
steps. It just works "properly" (although with a meaningless result due to
the long interval between savings) on my next test, where the .trr file is
saved just every 100 time steps. :(

This seems to happen on simulations consisting of 100,000 time steps,
250,000 time steps or 500,000 time steps in total: unsure about what would
happen with a million time steps yet.

Does anybody has an idea of what is going on, and how can I try to solve it?

Thanks a lot in advance!  :)

Jones