[Wien] Segmentation fault in pairhess with constrains

Oleg Rubel rubelo at tbh.net
Tue Dec 8 19:49:35 CET 2009


Dear Laurence,

thank you for the reply. I downloaded the most recent version of pairhess and compared to that I used with 'diff' command. I did not find any differences except for the compiler options. Anyway, I recompiled the just downloaded pairhess with the options suggested and tried it. The error persists:

    [oleg at feynman t-Se]$ /home/oleg/temp/pairhess/pairhess pairhess.def
     case.inM present and used for constrains
     Average Hessian Eigenvalue      0.0 mRyd/au^2, Frequency    0.00 cm-1
    forrtl: severe (174): SIGSEGV, segmentation fault occurred
    Image              PC                Routine            Line        Source
    pairhess           0000000000407CAF  MAIN__                    587  main.f
    pairhess           000000000040385C  Unknown               Unknown  Unknown
    libc.so.6          0000003F9161D8B4  Unknown               Unknown  Unknown
    pairhess           0000000000403769  Unknown               Unknown  Unknown

The segfault points to the following line of main.f:

    write(*,704)'Min & Max of Eigenvalues, mRyd/au^2 ',1000*W2(norigs+1),1000*w2(if1)

I also noticed a strange 'Inf' and 'NaN' in case.outputpair:

    [oleg at feynman t-Se]$ tail -n 20 t-Se.outputpair
     Hxx, Hyy, Hzz atom   6   0.7010   0.7010   0.7010 |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   7   0.7010   0.7010   0.7010 |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   8   0.7010   0.7010   0.7010 |  Fixed Fixed Fixed   Mult   4 Z    34.0


     Rescaling Hessian by     -Inf
     Average Eigenvalues       NaN mRyd/au^2
     Diagonal Elements of scaled Hessian
     *************************************
     Hxx, Hyy, Hzz atom   1-Infinity-Infinity-Infinity |  Free  Free  Free    Mult   4 Z     3.0
     Hxx, Hyy, Hzz atom   2-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   3-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   4-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   5-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   2 Z    34.0
     Hxx, Hyy, Hzz atom   6-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   7-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0
     Hxx, Hyy, Hzz atom   8-Infinity-Infinity-Infinity |  Fixed Fixed Fixed   Mult   4 Z    34.0


     ********* Eigenmode Listing *********

Thank you once again.

Oleg


>>> Laurence Marks <L-marks at northwestern.edu> 12/08/09 12:23 PM >>>
I know this bug and I believe it has been cured in the version that is
on the web -- please download just pairhess from the web page (I am
currently on travel so cannot do this simply).

2009/12/8 Oleg Rubel <rubelo at tbh.net>:
> Dear Wien2k Users and Developers,
>
> I attempt to model a cleaved surface of trigonal selenium terminated by Li atoms using WIEN2k_09.2 (Release 29/9/2009). In the calculation I would like to constrain the position of Se-atoms and let Li-atoms to relax free. I prepared my case.inM accordingly (atom 1 - Li, atoms 2-8 - Se):
>
>    [oleg at feynman t-Se]$ cat t-Se.inM
>    PORT 2.00 0.35      # PORT/NEWT;  tolf, Initial Trust Radius
>    1.0 1.0 1.0 1.0   #Atom    1 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    2 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    3 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    4 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    5 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    6 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    7 Generated by pairhess
>    0.0 0.0 0.0 1.0   #Atom    8 Generated by pairhess
>
> Then I run pairhess and get segmentation fault (the error file does not contain any useful information.)
>
>    [oleg at feynman t-Se]$ x pairhess
>     case.inM present and used for constrains
>    forrtl: severe (174): SIGSEGV, segmentation fault occurred
>    Image              PC                Routine            Line        Source
>    pairhess           0000000000413299  Unknown               Unknown  Unknown
>    pairhess           0000000000409A9C  Unknown               Unknown  Unknown
>    libc.so.6          0000003F9161D8B4  Unknown               Unknown  Unknown
>    pairhess           00000000004099A9  Unknown               Unknown  Unknown
>    0.096u 0.177s 0:00.91 28.5%     0+0k 0+0io 4pf+0w
>    error: command   /home/oleg/WIEN2k_09.2.mkl_10_IB_MVAPICH2/pairhess pairhess.def   failed
>
> Then I tried an older version of WIEN2k_08.3 (Release 18/9/2008), and it works
>
>    [oleg at feynman t-Se]$ x pairhess
>     PORT           0
>     case.inM present and used for constrains
>     Max min of Eigenvalues    1.41645881505472        1.58354118615782
>     Max min of I-H Eigen    -0.583541186157820      -0.416458815054722
>     Free elements condition number    1.11795780570241
>     Condition number  0.15835D+01 for     3 variables; Cond/sqrt(N)= 0.91426D+00
>     Check .minpair, the estimate, and output in
>     /home/oleg/PROJECTS/TRIGONAL-Se.DEF-POTENTIAL/WIEN2K/t-Se.SURFACE-UNSTRAINED/t-
>     S
>     PairHess END
>    0.001u 0.003s 0:00.10 0.0%      0+0k 0+0io 4pf+0w
>
> I should mention that without constrains pairhess of WIEN2k_09.2 (Release 29/9/2009) runs error-free. I know that pairhess has recently undergone some major changes. Could it be that a bug was introduced? I enclose my case.struct and case.inM just in case of anyone would like to reproduce the error.
>
> Thank you very much in advance.
>
> Oleg Rubel
>
> --
> Oleg Rubel, PhD
> Scientist
> Thunder Bay Regional Research Institute
> 290 Munro St
> Thunder Bay, ON
> P7A  7T1, Canada
> Phone: +1-807-7663350
> Fax: +1-807-3441948
> E-mail: rubelo at tbh.net
> Homepage: http://www.tbrri.com/~orubel/
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
>



-- 
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2220 N Campus Drive
Northwestern University
Evanston, IL 60208, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
email: L-marks at northwestern dot edu
Web: www.numis.northwestern.edu
Chair, Commission on Electron Crystallography of IUCR
www.numis.northwestern.edu/
Electron crystallography is the branch of science that uses electron
scattering and imaging to study the structure of matter.
_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien



More information about the Wien mailing list