[Wien] Segmentation fault in lapw0 -eece

Gavin Abo gsabo at crimson.ua.edu
Thu Sep 5 08:44:29 CEST 2013


I wonder if this could be caused by the fftpack. The fftpack was fine in 
11.1, but it was broken in 12.1 and required fixes to fft_modules.F and 
vresp.F. The fft_modules.F looks fixed in 13.1. So far I haven't notice 
any problems, but maybe a case has been found that still needs the patch 
to vresp.F 
(http://zeus.theochem.tuwien.ac.at/pipermail/wien/2012-November/017911.html).

So my suggestion would be to try the patch or compile using serial 
settings with a fftw library (e.g., search the mail archive for DFFTW3).

On 9/4/2013 10:29 PM, 西村 真一 wrote:
> Dear Wien2k users,
>
> I met problem in -eece calcualtion.
> The operation I did was as follow;
> 1. init
> 2. runsp
> 3. save -d xxx
> 4. edit case.ineece (copied from SRC_templates)
> 5. runsp -eece
>
> After the 2nd lapw2, "x lapw0 -eece" failed.
> The output is:
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image              PC                Routine            Line        Source
> lapw0              00000000004804A0  MAIN__                    695  lapw0.F
> lapw0              000000000040388C  Unknown               Unknown  Unknown
> libc.so.6          00002B38BDBC3C8D  Unknown               Unknown  Unknown
> lapw0              0000000000403789  Unknown               Unknown  Unknown
> 0.036u 0.004s 0:00.04 75.0%	0+0k 0+40io 0pf+0w
>
> The line #695 of lapw0.F for ver.12.1 (#704 for ver.13.1) is
> 	clmsp(j,lm1,jatom,1)=clmsp(j,lm1,jatom,1)*sqfp.
>
> This problem does not occur in WIEN2k ver.11.1a, while ver.12.1 and ver.13.1 gave the same error.
> The compiler and the linked libraries are same. (Intel Composer XE 2013.5.192)
>
> Thanks in advance,
>
> Shinichi Nishimura


More information about the Wien mailing list