[Wien] rewind case.boyd1 error (mixer, qmix9.f)
Laurence Marks
laurence.marks at gmail.com
Wed Jan 15 13:17:55 CET 2025
Ok, now we know where the issue is, we don't yet know why.
One possibility is that a NaN or something similarly odd is being written
in one cycle and cannot later be read. Please send the case.scf & case.scfm
for a problem that died to me personal email. I can then look for anything
odd in the iterations. I will send later (after more coffee and some
breakfast) a couple of code tweaks to help find out more about the issue.
Unfortunately I cannot think of a way around this except for deleting all
the *broyd* files. I think it is better to try and understand why this is
happening.
Please also run
ldd $WIENROOT/mixer
and send me the results.
___
Emeritus Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering, Northwestern University
www.numis.northwestern.edu
"Research is to see what everybody else has seen, and to think what nobody
else has thought" Albert Szent-Györgyi
On Wed, Jan 15, 2025, 01:56 Sergeev Gregory <sgregory at live.ru> wrote:
> I tried this,
> in file case.scfm I got "Input file corruption"...
>
> If we assume that this is a problem with the file system, does it make
> sense to write the file twice (say *broyd.1backup) and if *broyd.1 cannot
> be read, switch to reading *broyd.1backup?
>
> About "different versions of ifort", I don't think so, because problem
> often is happend nth-cycle.
>
> - Gregory
>
> ------------------------------
> *От:* Wien <wien-bounces at zeus.theochem.tuwien.ac.at> от имени Laurence
> Marks <laurence.marks at gmail.com>
> *Отправлено:* 14 января 2025 г. 17:04
> *Кому:* A Mailing list for WIEN2k users <wien at zeus.theochem.tuwien.ac.at>
> *Тема:* Re: [Wien] rewind case.boyd1 error (mixer, qmix9.f)
>
> One possibility is that you have different versions of ifort (if that is
> what you used) on the node where you compiled the code and one or more of
> those used to run it. Check with your sys admin about this. You may be able
> to use static linking to overcome this if the cluster is heterogeneous.
>
> ---
> Emeritus Professor Laurence Marks (Laurie)
> www.numis.northwestern.edu
> https://scholar.google.com/citations?user=zmHhI9gAAAAJ&hl=en
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought" Albert Szent-Györgyi
>
> On Tue, Jan 14, 2025, 05:03 Laurence Marks <laurence.marks at gmail.com>
> wrote:
>
> I suspect the file got corrupted. Please change those lines to
>
> 9901 dmix_tmp =1.D0
> BetaS_tmp=1.D0
> Write(21,*)'Input file corruption'
> 9902 rewind(31)
>
> This won't cure the issue, but will be a diagnostic. I will send something
> more later.
> Explanation: just above this the code is reading in some stored
> parameters. If I am right this has gone wrong, and it has jumped to 9901.
> Why the rewind is failing I don't understand. My only guess is a right
> issue or some file corruption'. The addition will test if my guess is right.
> ___
> Emeritus Professor Laurence Marks (Laurie)
> Department of Materials Science and Engineering, Northwestern University
> www.numis.northwestern.edu
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought" Albert Szent-Györgyi
>
> On Tue, Jan 14, 2025, 04:45 Sergeev Gregory <sgregory at live.ru> wrote:
>
> Thank you for quick reply,
> I use version 24.1, 126 line is:
>
> 9901 dmix_tmp =1.D0
> BetaS_tmp=1.D0
> 9902 rewind(31) <---- 126 line
> !
> ! Cleanup ATLIM without atoms
>
> Hope it will help to catch problem.
>
> - Gregory
>
> ------------------------------
> *От:* Wien <wien-bounces at zeus.theochem.tuwien.ac.at> от имени Laurence
> Marks <laurence.marks at gmail.com>
> *Отправлено:* 14 января 2025 г. 13:19
> *Кому:* A Mailing list for WIEN2k users <wien at zeus.theochem.tuwien.ac.at>
> *Тема:* Re: [Wien] rewind case.boyd1 error (mixer, qmix9.f)
>
> You have not provided enough information. It looks like you are using an
> old version:
>
> What version of Wien2k?
> What release of the mixer (stated in case.scfm)?
> What compiler?
> What is line 126 of qmix9.f in your version?
>
> You can't recreate the file. It may have got corrupted when a job ended or
> for another reason such as a bad calculation.
>
> ___
> Emeritus Professor Laurence Marks (Laurie)
> Department of Materials Science and Engineering, Northwestern University
> www.numis.northwestern.edu
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought" Albert Szent-Györgyi
>
> On Tue, Jan 14, 2025, 02:13 Sergeev Gregory <sgregory at live.ru> wrote:
>
> Dear developers,
> I do my calculations on hpc with slurm system with lustre and from time to
> time my task crash (approxymatly each 5th task) with error:
>
> forrtl: Invalid argument
> forrtl: severe (20): REWIND error, unit 31, file .../case.broyd1
> Image PC Routine Line Source
>
> mixer 0000000000487B4C Unknown Unknown Unknown
> mixer 00000000004AD18C Unknown Unknown Unknown
> mixer 0000000000458151 qmix9_ 126 qmix9.f
> mixer 0000000000419B5B MAIN__ 1600 mixer.F
> mixer 0000000000403A7E Unknown Unknown Unknown
> libc-2.12.so 0000003EC121ED1D __libc_start_main Unknown
> Unknown
> mixer 0000000000403929 Unknown Unknown
> Unknown
>
> Is there a solution to this problem?
> Can I delete the file and create a new one with the same name without
> breaking the functionality in case of rewind's iostat != 0?
>
> This problem is very annoying, I will be grateful if you help me find a
> solution.
>
> - Gregory
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20250115/b5536dc5/attachment.htm>
More information about the Wien
mailing list