[Wien] small errors in tetra.f
I've found 3 problems in tetra.f, which are fixed in the attached file. a) When you have only 1 k-point, tetra switches correctly to Gauss-broadening, but used a broadening parameter from case.in3 (3rd line) without checking if it is non-zero. Zero broadening leads to NaNs in the DOS. b) The array TEXT should be allocated with (0:...) c) format 1030 changed for correct headers in the dos files. Regards -- P.Blaha -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWIEN2k: http://www.wien2k.at WWW: http://www.imc.tuwien.ac.at/staff/tc_group_e.php -- PROGRAM TETRA ! ! D E N S I T YO FS T A T E S ! ACTUAL VERSION with BLOECHL SCHEME ! use reallocate INCLUDE 'param.inc' CHARACTER *7 A7 CHARACTER *9 A9,B9 CHARACTER *20 TITLE CHARACTER*96,allocatable :: TEXT(:) CHARACTER *70 SYSTEM CHARACTER*11 FORM,STATUS CHARACTER*180 FNAME,fname1,fname2 CHARACTER*80 errfn character*2 fnameupdn,fnum2 character*1 fnum1 ,abrsw ! real*4,allocatable :: RNUMB(:,:), DENSTY(:,:),dsum(:) real*4 fermden(MG) COMMON /NCOUNT/ NNOC real*4,pointer :: EBS(:,:), FC(:,:,:) CHARACTER *14 DOSTYP(MG) DIMENSION IDOS(MG,2) real*4,allocatable :: QTL(:,:),xqtl(:,:) real*4,allocatable :: ehelp(:),denbr(:,:) integer,allocatable :: isumdos(:,:) logical rxes,rxesw,enefile common /rxescom/ rxes,rxesw ! DATA PI /3.141592654/, EEF /0.0/ ! ! call getarg(2,fname) rxes=.false. rxesw=.false. enefile=.false. CALL GTFNAM(fname,ERRFN) ! if(fname.eq.' ') call getarg(1,fname) OPEN(1,FILE=fname,STATUS='OLD',ERR=8000) 8003 READ(1,*,END=8001) IUNIT,FNAME,STATUS,FORM,IRECL OPEN(IUNIT,FILE=FNAME,STATUS=STATUS,FORM=FORM,ERR=8002) if(iunit.eq.7) then do i=180,5,-1 if(fname(i:i).eq.'1') then nnchar=i-1 fname1(1:i-1)=fname(1:i-1) fnameupdn='' GOTO 8003 else if(fname(i:i).eq.'p') then nnchar=i-3 fname1(1:i-3)=fname(1:i-3) fnameupdn='up' GOTO 8003 else if(fname(i:i).eq.'n') then nnchar=i-3 fname1(1:i-3)=fname(1:i-3) fnameupdn='dn' GOTO 8003 endif enddo endif if(iunit.eq.108) then enefile=.true. kspin=2 if(fnameupdn.eq.'') kspin=1 endif if(iunit.eq.8) then rxes=.true. read(8,'(a)')system write(6,'(RXES spectroscopy for,a )') system endif if(iunit.eq.9) then rxesw=.true. write(6,'(RXES spectroscopy weight file case.rxes calculated)') endif GOTO 8003 8000 WRITE(*,*) ' ERROR IN OPENING TETRA.DEF ' STOP 'TETRA.DEF' 8002 WRITE(*,*) ' ERROR IN OPENING UNIT:',IUNIT WRITE(*,*) ' FILENAME: ',FNAME,' STATUS: ',STATUS,' FORM:',FORM STOP 'OPEN FAILED' 8001 CONTINUE CLOSE (1) !-- READ(5,1) TITLE IAV=0 NPRINT=1 ! MI=1000 eef=999. ! EEF=0. ! NST number of k points ! IAV 0 NO PLOT OTHERWISE DOS AVERAGED OVER IAV VALUES ! NPRINT 0 NO DOS PRINTED ! 1 DOS AND INTEGRATED DOS PRINTED ! NYMINLOWER BAND INDEX ! NYMAXUPPER BAND INDEX !
[Wien] octave3
Dear WIEN2k team, Problem with octave sshift: t1=loadstruct(GaAs.struct) t2=makeconventional(t1) t3=sshift(t2,1,[0.0 0.0 0.1]) sshift makes no operation and t3=t2. The full content of t2 and t3 are written out on the screen, so t3=t2 is easily seen. On the screen a diagnosis appears: octave:4 t3=sshift(t2,1,[0.0 0.0 0.1]) sh: 1: tcsh: not found cp: cannot stat ‘tmpsave.struct_ncmsym’: No such file or directory sh: 1: tcsh: not found cp: cannot stat ‘tmpsave.struct_ncmsym’: No such file or directory smultatom: no spgroup specified, sym. op. are generated showequivalent: no spgroup specified, sym. op. are generated warning: showequivalent: some elements in list of return values are undefined My WIEN2k is successfully installed with .bashrc. How can one get sshift going? Regards, John Rundgren ___ Wien mailing list Wien@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
Re: [Wien] octave3
Apparently you have installed a csh (needed for all our *_lapw scripts), but not a tcsh. In most modern systems these are anyway the same. Solutions: a) Install a tcsh on your system b) fix the source code: In $WIENROOT/SRC_structeditor/bin search for all lines with tcsh: grep tcsh *.m It gives a list where tcsh occurs: savestruct.m:system(tcsh -c 'xncm ncmsymmetry -f tmpsave -nomm error.ncmsym' ); showstruct.m:# system(tcsh -c 'struct2dx -nomm -new -f s error.struct2dx', 1 , async); showstruct.m: system(tcsh -c 'xcrysden --wien_struct s.struct error.xcrysden', 1); showstruct.m:# system(tcsh -c 'xcrysden --wien_struct s.struct error.xcrysden', 1 , async); showstruct.m:# system(tcsh -c 'VESTA $PWD/s.struct error.VESTA', 1 , async); s You can ignore the commented lines, but need to edit savestruct.m and showstruct.m and replace tcsh by csh Please let me know if it works. I'll probably modify the sources for the next release. Regards On 09/24/2014 12:15 PM, John Rundgren wrote: Dear WIEN2k team, Problem with octave sshift: t1=loadstruct(GaAs.struct) t2=makeconventional(t1) t3=sshift(t2,1,[0.0 0.0 0.1]) sshift makes no operation and t3=t2. The full content of t2 and t3 are written out on the screen, so t3=t2 is easily seen. On the screen a diagnosis appears: octave:4 t3=sshift(t2,1,[0.0 0.0 0.1]) sh: 1: tcsh: not found cp: cannot stat ‘tmpsave.struct_ncmsym’: No such file or directory sh: 1: tcsh: not found cp: cannot stat ‘tmpsave.struct_ncmsym’: No such file or directory smultatom: no spgroup specified, sym. op. are generated showequivalent: no spgroup specified, sym. op. are generated warning: showequivalent: some elements in list of return values are undefined My WIEN2k is successfully installed with .bashrc. How can one get sshift going? Regards, John Rundgren ___ Wien mailing list Wien@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 -- P.Blaha -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWIEN2k: http://www.wien2k.at WWW: http://www.imc.tuwien.ac.at/staff/tc_group_e.php -- ___ Wien mailing list Wien@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
Re: [Wien] wien2wannier with wien2k.14.1
It is a matter of taste how one wants to do it. The UTF-8 character set has nice features with special characters, for instance displays z^2 really with a superscript 2. On the other hand, when using putty from home, I had the same problem as you and unreadable characters appear as output and sometimes the messages are completely scrambled. It should not affect basic functionality (except when you cannot read some help-instruction). On 09/24/2014 03:32 AM, Zhu, Jianxin wrote: All, Sorry for disturbing you this afternoon. After a few more times of struggling, I pinned down the root cause to be with the setting of xterm under x11. With the correction, the problem disappears. Of course, we still need to setenv LANG en_US.UTF-8 on the linux cluster if the UTF-8 is not set as default, if we remotely log onto it. Cheers, Jianxin On 9/23/14 12:22 PM, Zhu, Jianxin jx...@lanl.gov wrote: Hi Elias, As you suggested earlier, I pulled this issue to the mailing list. After I downloaded wien2k.14.1 integrated with wien2wannier, even before compilation/installation, I noticed many strange characters displayed inside the wien2wannier90 script files, for example, near the end of init_w2w_lapw OPTIONS: -b - batch mode (do not ask questions) -numk - number of k-points (default: 100; can also use â~@~\Nx Ny Nzâ~@~]) -emin,-emax - select bands by energy (needs â~@~Xoutput1â~@~Y file; careful!) -all X Y- synonym for â~@~X-emin X -emax Yâ~@~Y -bands I J - select bands directly (more robust) -proj - specification of initial projections (â~@~\Amnâ~@~]) -w90- Wannier90 options to set â~@~XOPTIONSâ~@~Y are only operative in batch mode, except â~@~X-w90â~@~Y, which accepts as arguments key-value pairs to be put into â~@~Xcase.winâ~@~Y. The sp/so switches are usually unnecessary (detected automatically). theend exit 0 I did some file type check (with the suggestion of Matt Jones at Univ. Buffalo) --- /wien2k.14.1] jxzhu% file --mime init_lapw init_lapw: text/x-shellscript; charset=us-ascii /wien2k.14.1] jxzhu% file --mime init_w2w_lapw init_w2w_lapw: text/x-shellscript; charset=utf-8 It looks the charset used for wien2wannier90 scripts is different than the rest of wien2k scripts. As such, I manually set setenv LANG en_US.UTF-8 in the login shell script file. Then I see the following --- OPTIONS: -b - batch mode (do not ask questions) -numk - number of k-points (default: 100; can also use âNx Ny Nzâ file; careful!) -all X Y- synonym for â -bands I J - select bands directly (more robust) -proj - specification of initial projections (âAmnâ are only operative in batch mode, except â, which accepts as arguments key-value pairs to be put into â. The sp/so switches are usually unnecessary (detected automatically). theend exit 0 With the manual setting of UTF-8, I still see characters like hat{a} and the item like -w90 is missing in the display. As such, I am not sure what I get in whatever way is correct. I tested this issue on my Mac OSX machine, but also saw the strange characters on my linux cluster. Kevin, I appreciate if you can help me check this on your mac osx machine. Cheers, Jianxin On 7/1/14 5:34 AM, Elias Assmann elias.assm...@gmail.com wrote: Dear wien2wannier Users! A new version of the package is available at http://www.ifp.tuwien.ac.at/forschung/arbeitsgruppen/cms/software-downlo a d/wien2wannier/. This is a minor update that fixes a few bugs, including one in Œwplot2xsf¹ that caused it to crash. -- Elias Assmann (TU Wien) Wien2Wannier: maximally localized Wannier functions from linearized augmented plane waves http://www.ifp.tuwien.ac.at/forschung/arbeitsgruppen/cms/software-downlo a d/wien2wannier/ ___ Wien mailing list Wien@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@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@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 -- P.Blaha -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWIEN2k: http://www.wien2k.at WWW: http://www.imc.tuwien.ac.at/staff/tc_group_e.php
Re: [Wien] wien2wannier with wien2k.14.1
On 09/24/2014 03:32 AM, Zhu, Jianxin wrote: Sorry for disturbing you this afternoon. It is for me to apologize for inflicting this problem on you (and on other wien2wannier users, I fear) through my choice to use UTF-8 characters in some wien2wannier messages. I chose to use those characters because they were nice for formatting the names of orbitals for the initial projections in `write_inwf' (like Peter's example `z²'), and because I thought that support for them should be pretty much universal on any modern system. Of course I was aware that there might be a few people it would fail for, but not that it would fail so badly. (You showed that not only are the Unicode characters displayed wrongly, but ASCII character between them are deleted: `-emin X -emax Y' [substituting apostrophes for the offending Unicode quote marks] becomes â!) I will shortly provide an ASCIIized version of wien2wannier, and avoid non-ASCII characters in future versions. (Of course, the Right Thing would be to make the programs aware of the locale settings and adjust the messages accordingly, as many tools do these days ...) After a few more times of struggling, I pinned down the root cause to be with the setting of xterm under x11. With the correction, the problem disappears. Of course, we still need to setenv LANG en_US.UTF-8 on the linux cluster if the UTF-8 is not set as default, if we remotely log onto it. On my machine, experimentation shows that it is a question of the locale settings in effect when the terminal is started. I.e., $ unset LANG $ xterm gives me a terminal that displays the characters wrongly, but just changing it in the current shell [old terminal] $ xterm [new xterm]$ unset LANG does not change the behavior. Also, xterm showed the problem while my standard terminal application does not. Elias ___ Wien mailing list Wien@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