[Wien] The TELNES errors in calculation Hf O23 edge indefective HfO2 supercell structure

Wed Aug 9 08:43:32 CEST 2006

Hello Mino,

while contemplating the possible benefits and risks of an error massage, I looked at your case.  There are several things to say :

* The case.elnes file in your tar-file contains a spectrum, but one that was written before the current calculation (i.e., it corresponds to a different case.innes file, and is not part of the calculation that crashed).  Can you confirm this?
* On my machine, your calculation finishes without errors or crashes.  ( I had to add a few lines to the input file to substitute for the missing case.qtl which you didn't send (they get a lot smaller when gzipped), but it's really the same thing.)  So, it seems the problem may be compiler related.
* The edge you are calculating is a transition from the 5p state.  This is NOT considered a core state by WIEN2k (it's at -50 eV or so), and while telnes2 seems to be able to do the calculation, you should check carefully that there are no orthogonality issues, as this isn't really what the program was meant to do (perhaps I will make the code give a warning for cases like this).
* Your crash occurs right at the end of the program, where an input file for the broadening program is prepared.  This is unfortunate - telnes2 has calculated the spectrum, but it crashes right before writing it to file ...

So, where do we take it from here?  I know exactly where your case crashes, but I cannot reproduce the crash, and I don't see why there should be a crash (it happens in a fairly trivial part of the code).
* If you have another fortran 90 compiler, could you give the other compiler a try?  Or try the current one with different settings (eg., -O0)?  You don't have to recompile all of wien2k, only the telnes2 program.
* If that's not possible, can you tell me which compiler and platform you use?  Or let me access your machine?  (If sending a login, reply to me only - not to the whole mailing list.)
* In the first e-mail, you mention that the O K edge also crashes sometimes.  Can you send me an example of that calculation?

Best regards,

Kevin Jorissen

Kevin Jorissen
University of Washington
Dept. of Physics
Box 351560
Seattle, WA-98195-1560
U.S.A.

phone  +1 206 543 3904
fax      +1 206 685 0635

e-mail kevin.jorissen at ua.ac.be
web page http://fraangelico.phys.washington.edu/~jorissen

________________________________

From: mino yang [mailto:yangmino at samsung.com]
Sent: Thu 8/3/2006 1:48
To: Jorissen Kevin
Subject: Re: Re: [Wien] The TELNES errors in calculation Hf O23 edge indefective HfO2 supercell structure

Dear Kevin
Thank you for the response.

I attache the necessary files except case.qtl file which are over 100Mbyte - it exceed my mail's capability.
And I did not compress the vacant file which have name but no contents, O byte.

As the previous mail, I'm trying to calculate Hf O23 edge of the 211HfO supercell. 
Actually I'm beginner. So it will be not sufficient for you inspection.

So don't mind demanding what you need more.

My best

Mino

the error massage when I ran "x telnes2"
------------------------------------------------------------
Input/Output Error 218: Recursive IO was detected

   In Procedure: writeaveragedenergyspectrum
        At Line: 68

      Statement: List-Directed WRITE
           Unit: 8
   Connected To: 211HfO.inb
           Form: Formatted
         Access: Sequential
Records Read   : 0
Records Written: 4

End of diagnostics

29.900u 0.650s 0:30.83 99.0%    0+0k 0+0io 175pf+0w
------------------------------------------------------------

A/ The person helping you can see easily what you're doing ;
B/ This person can easily reproduce your calculation, without having to run it from the start (especially when doing supercells!).  Eg., when it's about telnes2, a look at telnes2.def will tell you which files will be of use to me.

------- Original Message -------
Sender : Jorissen Kevin<Kevin.Jorissen at ua.ac.be>
Date   : 2006-08-03 16:07
Title  : Re: [Wien] The TELNES errors in calculation Hf O23 edge in     defective
 HfO2 supercell structure

Hello ? (seems your name got cut off?)

Thanks for your feedback.

1/  You did the right thing : all inequivalent spectra need to be calculated separately, then summed.  No shortcuts here.  The ATOMS option sums equivalent atoms - not inequivalent atoms.

2/  Thanks for including case.struct, but there's still not enough information here for me to diagnose your error message.  Please send details, including files, to my private account (the wien ML blocks larger attachments).  As a general rule when asking for help, send files such that A/ The person helping you can see easily what you're doing ; B/ This person can easily reproduce your calculation, without having to run it from the start (especially when doing supercells!).  Eg., when it's about telnes2, a look at telnes2.def will tell you which files will be of use to me.

Thanks,

Kevin Jorissen

Kevin Jorissen
University of Washington
Dept. of Physics
Box 351560
Seattle, WA-98195-1560
U.S.A.

phone  +1 206 543 3904
fax      +1 206 685 0635

e-mail kevin.jorissen at ua.ac.be
web page http://fraangelico.phys.washington.edu/~jorissen

________________________________

From: wien-bounces at zeus.theochem.tuwien.ac.at on behalf of ???
Sent: Wed 8/2/2006 20:58
To: wien at zeus.theochem.tuwien.ac.at
Subject: [Wien] The TELNES errors in calculation Hf O23 edge in defective HfO2 supercell structure

Dear wien user.

I installed wien2k_05 version to Linux cluster which is operated by linuxif8.

I made up monoclinic HfO2 and its defective supercell structure, and got their Density of states.

Actually I would like to get their ELNES peak. I upgraded TELNES.2 program.

I tried to get Hf O23 edge (n=5, l=1) and O K edge of the defective supercell structure (2*1*1 HfO2 supercell with a vacancy in a Oxygen site).

There was always errors for the peak of Hf O23 edge and sometimes for the peak of O K edge.

I want to hear some advice about two below.

1. As it is supercell, I had to calculate every peak of uneqivalent O K edge and summed them. I'm not sure it is right method to get ELNES peak of O K edge. I saw "ATOMS" option in manual for the unequivalent atoms. Do I refer to it for my case?

2. Here is the error massage when I run telnes2 program to get the peak of Hf O23 edge. What should I do for the next step.
cat telnes2.error
Error while writing output in WriteAveragedSpectrum

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20060809/a6ec4e6e/attachment.html