[gmx-users] segmentation fault in g_anaeig
InonShar at post.TAU.ac.IL
Mon Aug 4 11:03:19 CEST 2008
Thanks, Tsjerk, you're absolutely right. I looked at the bugzilla
reports about g_anaeig. Bug 30 was already fixed on my version of the
source (pre-compilation) code. Bug 198 might be still relevant since
it's status is listed as NEW and it hasn't been resolved yet. The
bugzilla listing is as follows:
I encountered a segmentation fault when running g_anaeig, and found
out that it occurred when reading the eigenvectors. Apparently,
someone reported a similar problem for GMX 3.3.1. However, I looked
for the differences between the versions in eigio.c, and found out
that the original 3.3.1 code worked fine for me (under the specified
platform at least).
The changes I introduced:diff eigio.c eigio.org.c
< /* eignr=NULL; */
< /* eigval=NULL; */
< /*eigvec=NULL; */
> eignr=NULL; eigval=NULL; eigvec=NULL;
I'm not sure I understand how to implement this change: Currently, in
the eigio.c file in my source code the three lines do appear. Should I
turn them into comment lines and re-compile?
For general reference, I'm running GROMACS 3.3.3 on an AMD 64-bit
quadrocore, running under Fedora 8 and using bash. Any further
relevant details will be given happily.
Quoting "Tsjerk Wassenaar" <tsjerkw at gmail.com>:
> Hi Inon,
> Please be as complete as you can when posting questions. You've
> omitted the (probably) most important piece of information: what
> version are you using? (You might also want to include platform,
> compiler, etc.). Note that just a few days ago it was already
> on this list that g_anaeig of version 3.3.3 gave a segfault.
> On Sun, Aug 3, 2008 at 2:57 PM, Inon Sharony
<InonShar at post.tau.ac.il> wrote:
>> I am preforming a normal mode analysis (NMA) in Pentane, using
>> following command sequence (all functions ending in *_d are so
>> they were compiled in double-precision into a directory where the
>> precision-functions already existed):
>> grompp_d em.mdp ;energy minimization
>> mdrun_d em.mdp
>> grompp_d nm.mdp ;write Hessian matrix
>> mdrun_d nm.mdp
>> when running this last command I got:
>> trn version: GMX_trn_file (double precision)
>> Read mass weighted average/minimum structure with 5 atoms from
>> Segmentation fault
>> Why do I get a segmentation fault (segfault)?
>> Could it be because the energy minimization only reached a maximal
>> 1E-02 kJ / mole nm instead of the recommended 1E-05? Or is it some
>> serious problem?
>> I looked at the eigenfreq.xvg and saw 2 zero frequencies, and 4
>> are nearly zero.
>> I've already issued "make distclean" before the re-compilation in
>> double-precision (see my
>> Also, I viewed the eigenvec.tpr file and, indeed, the coordinates
>> double-precision format (i.e., five significant digits to the
right of the
>> decimal point are non-zero).
>> Any and all assistance would be greatly appreciated.
>> Enjoy your weekend,
>> Inon Sharony
>> ×× ×× ×©×¨×× ×
>> Please consider your environmental responsibility before printing
>> gmx-users mailing list gmx-users at gromacs.org
>> Please search the archive at http://www.gromacs.org/search before
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> Tsjerk A. Wassenaar, Ph.D.
> Junior UD (post-doc)
> Biomolecular NMR, Bijvoet Center
> Utrecht University
> Padualaan 8
> 3584 CH Utrecht
> The Netherlands
> P: +31-30-2539931
> F: +31-30-2537623
×× ×× ×©×¨×× ×
Please consider your environmental responsibility before printing
More information about the gromacs.org_gmx-users