[gmx-users] FW: error after 1ns of md run during production

[Mark Abraham]

On Wed, Mar 12, 2014 at 4:05 AM, Sucharita Dey <csisd at nus.edu.sg> wrote:

> I have done the subscription, hope my email will be accepted.
>
> Sucharita Dey,PhD, Research Fellow, Cancer Science Institute of Singapore,
> National University of Singapore. 14, Medical Drive,  #12-01, Singapore
> 117599. Tel: (65) 90355220, email: csisd at nus.edu.sg
> ________________________________________
> From: Sucharita Dey
> Sent: Wednesday, March 12, 2014 11:02 AM
> To: gromacs.org_gmx-users at maillist.sys.kth.se
> Subject: error after 1ns of md run during production
>
> Hello gmx-users,
>
> I am running gromacs in a cluster using mpi code, version is 4.5.5, getting
> the below mentioned error after 1ns run of the production run, before that
> I
> have successfully done 50 ps of equilibration. And I checked it is fine. Is
> the error because of the version? Please help.
>

It could be a bug that is resolved after 4.5.5, but nobody has enough
information to know. You'll be far better off getting 4.6.5 installed.
Also, 50ps is a rather quick equilibration, so instability leading to
http://www.gromacs.org/Documentation/Terminology/Blowing_Up is possible,
depending what you're doing.

Following is taken from the log file giving the error:
>
> Reading file /hpctmp/csisd/UTX_M//new/UTX_M-final.tpr, VERSION 4.5.5
> (single
> precision)
>
> Will use 16 particle-particle and 8 PME only nodes
> This is a guess, check the performance at the end of the log file
> Making 2D domain decomposition 8 x 2 x 1
>
> WARNING: This run will generate roughly 7742 Mb of data
>
> starting mdrun 'Protein in water'
> 25000000 steps,  50000.0 ps.
>
> Step 685356, time 1370.71 (ps)  LINCS WARNING
> relative constraint deviation after LINCS:
> rms 761119.299726, max 9567121.000000 (between atoms 2442 and 2443)
> bonds that rotated more than 30 degrees:
>  atom 1 atom 2  angle  previous, current, constraint length
>    2442   2443   90.0    0.1090 1042816.2500      0.1090
> Wrote pdb files with previous and current coordinates
> [cheetah-c59:mpi_rank_7][error_sighandler] Caught error: Segmentation fault
> (signal 11)
> [cheetah-c59:mpi_rank_6][error_sighandler] Caught error: Segmentation fault
> (signal 11)
>
> -------------------------------------------------------
> Program mdrun_mpi, VERSION 4.5.5
> Source code file: pme.c, line: 538
>
> Fatal error:
> 3 particles communicated to PME node 2 are more than 2/3 times the cut-off
> out of the domain decomposition cell of their charge group in dimension x.
> This usually means that your system is not well equilibrated.
> For more information and tips for troubleshooting, please check the GROMACS
> website at http://www.gromacs.org/Documentation/Errors
> -------------------------------------------------------
>
> Please note,  just to check, I ran the same md for 5ns in my PC which has
> gromacs 4.6.3, and it is running till now uninterrupted (already covered
> 3ns)
>

Ill-conditioned numerical models do not have reliable behaviour. Better
equilibration is almost certain to resolve your issues.

Mark


>
> Best,
> -Sucharita
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>