AW: Problem with NAMD 2.8 (but not in 2.7) : Endless loop repeating "The last position output ..."

From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Thu Jun 09 2011 - 00:44:01 CDT

Hi David,

Now I saw that you have dcdfrequency maybe set to 1 or not set, so he writes
the coordinates to dcd every step which is horrible for performance reason,
try to change this to a more reasonable value, also the veldcd if you have
and other files.

Best regards

Norman Geist.

-----Ursprüngliche Nachricht-----
Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im Auftrag
von David McGiven
Gesendet: Mittwoch, 8. Juni 2011 16:01
An: namd-l
Betreff: namd-l: Problem with NAMD 2.8 (but not in 2.7) : Endless loop
repeating "The last position output ..."

Dear NAMD Users,

I am running a NAMD Simulation taking as the starting point a
previously simulated data-file. The input does the following :

- It minimizes the complex for 100 ps with all the complex constraint
- Then it takes out the constraint and run for 50.000.000 steps

The minimization is done ok.

But then when the simulation starts (at step 7.000.000) the only
output I get is :

The last position output (seq=7000055) takes 0.019 seconds, 1453.703
MB of memory in use
The last position output (seq=7000056) takes 0.008 seconds, 1453.703
MB of memory in use
The last position output (seq=7000057) takes 0.019 seconds, 1453.703
MB of memory in use
[...]

Repeated endlessly.

The .xst and .dcd files grow some bytes every minute, but nothing
relevant. It’s like it gets stucked at this point.

This happens when running with NAMD 2.8b1 and 2.8 (final) no matter
how many processors I use.

This doesn’t happen with NAMD 2.7b1 no matter how many processors I use.

The problem is not random. It happens every time I run the namd command.

To reduce the sources of possible errors, I am running always with :

namd2 input.namd > output.log

No multicore, no network, no command line flags.

Also, this happens either restarting from .restart files or starting a
new simulation.

I am not attaching the input files since they’re ~40 Mb but can place
them online if needed.

Best Regards

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:57:16 CST