Re: Segmentation fault error with 'nan' pressure and energy

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Tue May 30 2017 - 17:14:16 CDT

Hi Sadegh,

Somehow, you've got a nan for a dihedral term. I've never seen this before, but the nan will propagate absurdly fast and poison all of the normal dynamics, leading to crashes very reliably. In alchemical simulations, it is normal to have lines in your log like this:

Info: ALCH: 13 ATOMS TO DISAPPEAR IN FINAL STATE
Info: ALCH: 13 ATOMS TO APPEAR IN FINAL STATE
Warning: ALCH: SUSPICIOUS BONDS BETWEEN INITIAL AND FINAL GROUPS WERE FOUND
Info: ALCH: 2 ANGLES LINKING INITIAL AND FINAL ATOMS DELETED
Info: ALCH: 12 DIHEDRALS LINKING INITIAL AND FINAL ATOMS DELETED

Do you have these lines? Or was this being run without turning the FEP/TI flags on? Minimization prior to FEP/TI needs to have these flags on too, otherwise the alchemify module won't remove angles and dihedrals that go across your appearing and disappearing groups.

-Josh

On 05/30/2017 03:56 PM, Sadegh Faramarzi Ganjabad wrote:
Richard,

Thanks for quick reply. Yes there are some identical atom coordinates. But I need it to be so, for alchemy transformation. Is there any way to deal with it?

Thanks and regards,
Sadegh

On Tue, May 30, 2017 at 5:41 PM, Richard Overstreet <roverst_at_g.clemson.edu<mailto:roverst_at_g.clemson.edu>> wrote:
I would check if atom coordinates are identical.

-Richard

On 05/30/2017 05:17 PM, Sadegh Faramarzi Ganjabad wrote:
Dear NAMD users,

I have seen couple of posts with the same topic, but still can't figure it out. I am trying to minimize a membrane protein system prier to run FEP/TI calculations. So, a part of my protein has a dual topology. I get 'Segmentation fault' with no further explanation. In my log file there are 'not a number, nan' values for pressure and energy. Here is a part of my log file

ETITLE: TS BOND ANGLE DIHED IMPRP ELECT VDW BOUNDARY MISC KINETIC TOTAL TEMP POTENTIAL TOTAL3 TEMPAVG PRESSURE GPRESSURE VOLUME PRESSAVG GPRESSAVG

ENERGY: 10 2846.5220 11429.2936 nan 185.1025 -94079.6680 9999999999.9999 0.0000 0.0000 0.0000 nan 0.0000 nan nan 0.0000 nan nan 450474.3936 nan nan

MINIMIZER SLOWLY MOVING 4 ATOMS WITH BAD CONTACTS DOWNHILL
PRESSURE: 11 nan nan nan nan nan nan nan nan nan
GPRESSURE: 11 nan nan nan nan nan nan nan nan nan
ENERGY: 11 3035.3700 11458.4002 nan 185.3481 -94074.5040 9999999999.9999 0.0000 0.0000 0.0000 nan 0.0000 nan nan 0.0000 nan nan 450474.3936 nan nan

MINIMIZER SLOWLY MOVING 4 ATOMS WITH BAD CONTACTS DOWNHILL
PRESSURE: 12 nan nan nan nan nan nan nan nan nan
GPRESSURE: 12 nan nan nan nan nan nan nan nan nan
ENERGY: 12 3266.8947 11491.0698 nan 185.6188 -94067.9283 9999999999.9999 0.0000 0.0000 0.0000 nan 0.0000 nan nan 0.0000 nan nan 450474.3936 nan nan

MINIMIZER STARTING CONJUGATE GRADIENT ALGORITHM
LINE MINIMIZER REDUCING GRADIENT FROM nan TO nan

There is nothing special about my configuration file. I have tested it on other systems and it works file. Does anybody has a clue to this issue?

Thanks,
Sadegh

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:20 CST