[Wien] QTL-related question - update

[Peter Blaha]

Sorry, I forgot to answer this as I havn`t solved it yet:

Yes, this is really a bug in the    qtl  program. It occurs only in parallel calculations
including the -telnes   switch  (no problems without this switch), which may explain that
nobody else detected it yet.

It needs some source code changes, similar to what has been done when reading the parallel
vector or energy files.


Am 01.12.2011 10:55, schrieb Ovidiu Cretu:
> Dear all,
>
>
> I don't know how to link this to the initial question, so I hope you see it.
>
> I have run the same calculations (same system, conditions, machine) in
> non-parallel mode and it works well. The relevant files in the folder are now
> the following:
> _________________________
> ls -l *.weigh*
>   6720 2011-12-01 01:34 [case].weigh
>      0 2011-11-30 17:55 [case].weighdn
>      0 2011-11-30 17:55 [case].weight
>      0 2011-11-30 17:55 [case].weightaver
> _________________________
>
> So it's either the way my machine does the parallel calculation (most likely) or
> a bug in the QTL code, which is very unlikely, as I'm the only one with the
> issue.
>
> Suggestions are welcome :)
>
>
> Regards,
>
> Ovidiu
>
> ----- Forwarded message from Ovidiu Cretu<ovidiu.cretu at ipcms.u-strasbg.fr>
> -----
>      Date: Tue, 29 Nov 2011 19:27:05 +0100
>      From: Ovidiu Cretu<ovidiu.cretu at ipcms.u-strasbg.fr>
> Reply-To: Ovidiu Cretu<ovidiu.cretu at ipcms.u-strasbg.fr>
>   Subject: QTL-related question
>        To: wien at zeus.theochem.tuwien.ac.at
>
> Dear all,
>
> I am running wien version 11.1 on a Dell Optiplex 755 running Fedora 12 (32bit),
> fortran compiler ifort 12.1.0 and math libraries MKL 10.3.7.
>
> The question concerns QTL, which I run in order to calculate some EELS spectra
> for a larger structure. The calculation is done, for both the SCF cycle and
> QTF, in parallel mode. The error that I get is the following:
>
> _________________________
> $ x qtl -telnes -c -p
> running QTL in parallel mode
> calculating QTL's from parallel vectors
> forrtl: severe (24): end-of-file during read, unit 26, file [case].weigh
> Image              PC        Routine            Line        Source
> qtl                081C9489  Unknown               Unknown  Unknown
> qtl                081C8205  Unknown               Unknown  Unknown
> qtl                0818A2D1  Unknown               Unknown  Unknown
> qtl                08159736  Unknown               Unknown  Unknown
> qtl                08158FA1  Unknown               Unknown  Unknown
> qtl                0816DE3E  Unknown               Unknown  Unknown
> qtl                0806826E  readw_                      6  readw.f
> qtl                0805ACC4  MAIN__                    493  qtlmain.f
> qtl                0804BAC4  Unknown               Unknown  Unknown
> libc.so.6          00BD2BB6  Unknown               Unknown  Unknown
> qtl                0804B9D1  Unknown               Unknown  Unknown
> 0.020u 0.024s 0:00.05 80.0%	0+0k 0+64io 0pf+0w
> _________________________
>
> The files that I have are the following:
>
> _________________________
> $ ls -l *.weigh*
>       0 2011-11-23 13:38 [case].weigh
>    8176 2011-11-29 16:28 [case].weigh_1
>    8168 2011-11-29 16:28 [case].weigh_2
>    4068 2011-11-29 16:28 [case].weigh_3
>       0 2011-11-23 13:38 [case].weighdn
>       0 2011-11-23 19:12 [case].weighdn_1
>       0 2011-11-23 19:12 [case].weighdn_2
>       0 2011-11-26 17:09 [case].weighdn_3
>   83690 2011-11-29 16:28 [case].weight
>       0 2011-11-23 13:38 [case].weightaver
> _________________________
>
> To me it looks like it has thee weigh_{1,2,3} files from/for a parallel
> calculation, but that QTL is looking for a single file, even though it is run
> in parallel mode. This .weigh file is empty, hence the error.
>
> Any ideas are more than welcome.
>
>
> Thanks in advance,
>
> Ovidiu
>

-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300             FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------