[Wien] initso in w2web and ABBRUCH: DIE EFG-MATRIX IST DIE NULLMATRIX !

Tue Nov 2 11:33:26 CET 2010

Dear Gerhard, 
I would also like to bring to your notice on the " ABBRUCH: DIE EFG-MATRIX
IST DIE NULLMATRIX"
I was running a spin polarized case with Fd-3m space group. The 
Command line was runsp_lapw -cc 0.000001 -in1ef -i 90 -p 
I used the w2web in initializing for spin orbit coupling. I went through all
the usual procedures of 
x symmetso 
x kgen etc 
ABBRUCH: DIE EFG-MATRIX IST DIE NULLMATRIX appears in the 3rd SCF cycle 

    cycle 3     (Tue Nov  2 15:17:22 IST 2010)  (88/97 to go)

>   lapw0 -p    (15:17:22) starting parallel lapw0 at Tue Nov  2 15:17:35
IST 2
10
-------- .machine0 : processors
running lapw0 in single mode
 ABBRUCH: DIE EFG-MATRIX IST DIE NULLMATRIX !
5.391u 0.131s 0:18.92 29.1%     0+0k 0+0io 17pf+0w
>   lapw1  -up -p       (15:17:59) starting parallel lapw1 at Tue Nov  2
15:18:
9 IST 2010
->  starting parallel LAPW1 jobs at Tue Nov  2 15:18:33 IST 2010
running LAPW1 in parallel mode (using .machines)
4 number_of_parallel_jobs
     ibnx90(265) 68.660u 1.459s 2:40.03 43.81%      0+0k 0+0io 0pf+0w
     ibnx90(265) 67.580u 1.717s 2:36.84 44.18%      0+0k 0+0io 0pf+0w
     ibnx78(265) 62.676u 1.277s 2:27.62 43.32%      0+0k 0+0io 0pf+0w
     ibnx78(265) 61.984u 1.225s 2:02.05 51.79%      0+0k 0+0io 0pf+0w
     ibnx90(1) 0.378u 0.034s 3.17 12.98%      0+0k 0+0io 0pf+0w
     ibnx90(1) 0.383u 0.019s 0.81 49.57%      0+0k 0+0io 0pf+0w
     ibnx90(1) 0.375u 0.022s 11.52 3.45%      0+0k 0+0io 0pf+0w

Thank you 
Suddhasattwa Ghosh 

-----Original Message-----
From: wien-bounces at zeus.theochem.tuwien.ac.at
[mailto:wien-bounces at zeus.theochem.tuwien.ac.at] On Behalf Of Gerhard Fecher
Sent: Tuesday, November 02, 2010 1:32 PM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] initso in w2web and ABBRUCH: DIE EFG-MATRIX IST DIE
NULLMATRIX !

Dear Peter,
soory it took me a while to test it
the new script works fine and the in2c is not longer changed after kgen.

There was recently a question on "ABBRUCH: DIE EFG-MATRIX IST DIE NULLMATRIX
!"
I realised that there are several different cases where it appeared in my
tests with so calculations:

usually it vanishes after the LM list is changed by mixer in the first scf
cycle
for not complex cases it hints that so was not switched on.

** case is not complex
1) In the first scf cycle of an SO calculation (only in the very first but
not if one restarts the calculation)
2) In all scf cycles if one forgets to switch on so
3) from the second cycle if one restarts a so calculation without so
in 2) or 3) the message is absent when one puts the LM list from .in2c (or
.in2_so)  into .in2

** case is  complex (usually only in the first cycle)
1) In the first scf cycle of an SO calculation (started after a non-SO
calculation)
2) In the first scf cycle if one forgets to switch on so (also not if one
continues afterwards with so)
but not if one restarts a so calculation without so

If one has a so calculation it is dangerous to use prepare new input files
form initialize calc. instead of initso
it may help if initso would overwrite the preprepared in2(c)_st files by the
in2(c)_so from initialize calc.

Question: Can anything  go wrong if one does not switch on so in the
so-symmetry ? 
                I guess it should only change the symmetry of the wave
functions and take longer due to the reduced symmetry
                but no additional SO splitting appears.

Ciao
Gerhard

====================================
Dr. Gerhard H. Fecher
Institut of Inorganic and Analytical Chemistry
Johannes Gutenberg - University
55099 Mainz
________________________________________
Von: wien-bounces at zeus.theochem.tuwien.ac.at
[wien-bounces at zeus.theochem.tuwien.ac.at]&quot; im Auftrag von &quot;Peter
Blaha [pblaha at theochem.tuwien.ac.at]
Gesendet: Freitag, 15. Oktober 2010 16:03
Bis: A Mailing list for WIEN2k users
Betreff: Re: [Wien] initso in w2web

Thanks for the report.

You may put the attached initso.pl into  $WIENROOT/SRC_w2web/htdocs/util
and I hope it fixes the problem.

Am 05.10.2010 18:55, schrieb Gerhard Fecher:
> may use the wrong in2c file if using initso_lapw from the w2web
webinterface
> when the direction is not 0 0 1
>
> With "Prepare new input files" the new  in2_so is copied to in2c
>
> Afterwards if one has an initially non complex case
> the in2c seems to be overwritten by x kgen by a different version.
>
> It can be easily tested for bcc Fe using 001 and 111 as quantization axes.
The LM in in2c will be the same after
> x kgen -so, but not the ones in the in2_so (I guess for bcc the 111 should
have always the 43, 63 components)
>
> It does not appear if the initial case was already complex (in2c_so)
>
> This does also not appear if one uses directly (not from w2web) the
initso_lapw script where the in2c is the same like in2_so.
>
> (I could not see from the pearlscript in the w2web utils folder where this
comes from)
>
> Workarround may be to call the "Prepare new input files" another time
after "x kgen -so"
>
> Ciao
> Gerhard
>
> ====================================
> Dr. Gerhard H. Fecher
> Institut of Inorganic and Analytical Chemistry
> Johannes Gutenberg - University
> 55099 Mainz
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

--

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW:
http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------
_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien