[Wien] LDA+U+SOC problem

Zhang Haijun hjzhang0428 at gmail.com
Thu Aug 2 23:01:24 CEST 2012 

On Wed, Aug 1, 2012 at 11:19 PM, Peter Blaha
<pblaha at theochem.tuwien.ac.at>wrote:

> i) If your case is non-magnetic, you should do the whole calculation with
>    run_lapw   and not with   runsp(_c)_lapw
>
      But I cannot just do LDA+U with run_lapw. Is it true?
      If it is true, how can I do LDA+U+SOC for cubic symmetry? This cubic
symmetry is very important.

>
> ii) Even when using   runsp_c_lapw, the program lapwdm checks the
> "symmetry" of
> a "magnetic SO" calculation, and it will stop, if you do not follow the
> suggestions
> of initso_lapw and break the symmetry properly by using the new struct
> file.
> Eventually, with runsp_c it may come out that the splitted atoms are still
> identical (because the moment is constrained to be zero), but you have to
> obey symmetry requirements.
>
    I don't understand why runsp_c_lapw is designed to use magnetic
symmetry. This does not make sense.
    For my case, the real symmetry has 48 operations, but the new structure
only has 16 operations. The symmety is too low.

     Please try to help me to do LDA+U+SOC with cubic symmetry. Thanks!


>
> I'd use   run_lapw.
>
>
>
> ------------------------------**------------------
> Von:
> Zhang Haijun <hjzhang0428 at gmail.com>
> Datum:
> 01.08.2012 10:15
> An:
> wien at zeus.theochem.tuwien.ac.**at <wien at zeus.theochem.tuwien.ac.at>
>
>
> Dear Wien2k developers,
>
>     But recently when I use wien2k to calculate LDA+U+SOC, I got some
> problems. Please help to fix it. Thank you very much!
>
>     The details are in the following:
>
>
> - I am running wien version 12.1 on a workstation with
> operating system centos 5, fortran compiler ifort 11.1 and math libraries
> mkl.
>
> - The purpose of my calculations is to get LDA+U+SOC.
>
> - I am running this case: these commands (eg. : run_lapw -p -so from
> command line, or w2web).
>
>     The compound is rock-salt structure SmS.
>     The case.struct, case.in1, case.inorb, case.indm(c), case.inso and
> other input files are attahced.
>     PBE is used, and rkmax=9. k-mesh=10x10x10
>
>    the commands:
>    1) initso_lapw
>       (because this coupond is nomagnetic, I keep the old symmetry here)
>    2) runsp_c_lapw -p -orb -so -NI -i 100
>
>
> - The program stops at this point, or produces suspicious output here
> ...  This is the content of the error file / the case.dayfile /
> STDOUT.  This is the bottom of the output file (like the part of
> case.output1 you show below).
>      1) error files: **  Error in Parallel LAPWDM
>
>      2) case.dayfile:
> Calculating SmS in /home/hjzhang/pub/wien2k/DATA/**SmS
> on centos1 with PID 6534
> using WIEN2k_12.1 (Release 22/7/2012) in /home/hjzhang/pub/wien2k/src
>
>
>     start       (Wed Aug  1 00:35:22 PDT 2012) with lapw0 (100/99 to go)
>
>     cycle 1     (Wed Aug  1 00:35:22 PDT 2012)  (100/99 to go)
>
> >   lapw0 -p    (00:35:22) starting parallel lapw0 at Wed Aug  1 00:35:22
> PDT 2012
> -------- .machine0 : processors
> running lapw0 in single mode
> 3.025u 0.519s 0:03.54 99.7%     0+0k 0+0io 0pf+0w
> >   lapw1  -up -p       (00:35:26) starting parallel lapw1 at Wed Aug  1
> 00:35:26 PDT 2012
> ->  starting parallel LAPW1 jobs at Wed Aug  1 00:35:26 PDT 2012
> running LAPW1 in parallel mode (using .machines)
> 1 number_of_parallel_jobs
>      centos1(47) 2.932u 0.697s 0:03.63 99.7%    0+0k 0+0io 0pf+0w
>    Summary of lapw1para:
>    centos1       k=47    user=2.932      wallclock=3.63
> 2.980u 1.111s 0:05.37 76.1%     0+0k 0+0io 0pf+0w
> copying vectorup to SmS.vectordn_1
> >   lapwso -up -orb -p  (00:35:32) running LAPWSO in parallel mode
>       centos1 4.009u 0.575s 0:04.58 99.7% 0+0k 0+0io 0pf+0w
>    Summary of lapwsopara:
>    centos1       user=4.009      wallclock=4.58
> 4.028u 0.860s 0:04.83 101.0%    0+0k 0+0io 0pf+0w
> >   lapw2 -c -up -so -p         (00:35:36) running LAPW2 in parallel mode
>       centos1 1.752u 0.285s 0:02.19 92.6% 0+0k 0+0io 0pf+0w
>    Summary of lapw2para:
>    centos1       user=1.752      wallclock=2.19
> 1.863u 0.746s 0:03.70 70.2%     0+0k 0+0io 0pf+0w
> >   lapwdm -up -p -so -c        (00:35:40) running LAPWDM in parallel mode
> **  LAPWDM crashed!
> 0.024u 0.243s 0:02.20 11.8%     0+0k 0+0io 0pf+0w
> error: command   /home/hjzhang/pub/wien2k/src/**lapwdmcpara -up -c -so
> uplapwdm.def   failed
>
> >   stop error
>
>   3) STDOUT:
>  LAPW0 END
>  LAPW1 END
> LAPWSO END
> LAPW2 - FERMI; weighs written
>  LAPW2 END
>  SUMPARA END
>
> >   stop error
>
>   4) the bottom of output:
>
>       7.1728964    7.1843841    7.2185287    7.2582704    7.2977652
>       7.3474635    7.4240100    7.4768003    7.4887500    7.4917268
>       7.5055530    7.5553705    7.5749512    7.5897230    7.6771047
>       7.6932632    7.7325403    7.7512731    7.7783675    7.8329630
>       7.9547802    7.9582511    7.9695529    7.9822074    7.9989245
>
>             0 EIGENVALUES BELOW THE ENERGY   -9.00000
>        **********************************************************
>
>        NUMBER OF K-POINTS:          47
>    ===> TOTAL CPU       TIME:      0.1 (INIT =      0.0 + K-POINTS =
>  0.1)
>       Maximum WALL clock time:    3.36647391319275
>       Maximum CPU time:           2.85000000000000
>
>
> - I have already tried the following things (bla bla bla) and they did
> or did not work, did or did not make any difference.
>
>   1) When I run initso_lapw, LDA+U+SOC can be done if I accept the new
> structure.
>       Because this coupond is nomagnetic, the old symmetry need to be
> held. if the new symmetry is accepted,
>       LDA+U_SOC cannot be done as I showed above.
>   2) When I run initso_lapw, LDA+SOC can be done if I keep the old
> structure.
>   3) LDA+U can be well done.
>
>     I have attahced all the input files. If possible, please give a try.
>
>
>
>
> You should also obey the following points:
>
> - Yes, I have browsed the archives AND READ THE USERS GUIDE and the
> FAQ pages Peter provides, but I couldn't solve my problem that way.
>
> - If suggestions from the mailing list actually solved your problem
> (or you solved it some other way), would you please share that with
> us?
>
>
>     I am looking forward to your suggestions! Thank you very much!
>
> --
> Best Wishes!
> Haijun Zhang
>
> --
> ------------------------------**-----------
> Peter Blaha
> Inst. Materials Chemistry, TU Vienna
> Getreidemarkt 9, A-1060 Vienna, Austria
> Tel: +43-1-5880115671
> Fax: +43-1-5880115698
> email: pblaha at theochem.tuwien.ac.at
> ------------------------------**-----------
> ______________________________**_________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.**at <Wien at zeus.theochem.tuwien.ac.at>
> http://zeus.theochem.tuwien.**ac.at/mailman/listinfo/wien<http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>
>



-- 
Best Wishes!
Haijun Zhang
====================
McCullough Building, Room 323
Stanford University
Stanford, CA 94305
Mobile: 650-8627320
Email: hjzhang0428 at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120802/b7841ce6/attachment-0001.htm>

More information about the Wien mailing list