[gmx-users] x2top segmentation fault in gromacs 5.0.x, but runs well for 4.0.x, 4.6.x

[Justin Lemkul]

On 3/27/15 7:47 AM, Hao Ren wrote:
> Hi there,
>
>
> I am trying a MD simulation for carbon nanotubes, but encountered a problem
> in the preparation of topology file.
>
> The coordinates and modified force filed can be found here:
> https://www.dropbox.com/s/qv4j33sivn56iey/test_x2top.tgz?dl=0
>
> I prepared the modified OPLS/AA force filed as described earlier in this
> list, post by Dr. Abhijit Kayal (
> https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-users/2014-July/090353.html
> ).
>
> A topology file can be generated by the g_x2top module from gromacs 4.0.7
> and 4.6.7 using the command line:
>
> g_x2top -f arm.gro -o topol.top -ff cnt_oplsaa
>
>
> However, I got a segmentation fault for gromacs 5.0.x, when I use the
> following command:
>
> gmx x2top -f arm.gro -o topol.top -ff cnt_oplsaa
>
> I have tried several OS/gcc combinations (centos 6.5/gcc-4.8.2,
> centos7/gcc-4.8.2, centos6.5/gcc-4.4.7, OSX10.9/homebrew-gcc-4.9.2) with
> fftw-3.3.4/gcc-4.8.2, all terminated with seg fault.
>
> I am not sure whether it comes from my invalid modification of force field
> or inappropriate compilation of gromacs or an internal bug of the gromacs
> code.
>
> Could anyone kindly help me to locate this problem? Or just test whether
> you can reproduce the error using your gromacs 5.0.x installation?
>

There is a bug in the code.  I found the offending commit and will file a 
Redmine issue.  In the meantime, just use an older version to generate the topology.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================