[Wien] Fwd: parallel case.vector and qtl
Peter Blaha
peter.blaha at tuwien.ac.at
Thu Dec 1 08:25:22 CET 2022
I do not have the full info and might have missed a few things, but:
Definitely there are a few bugs related to "qtl" in wien2k_21.
One is related to the qtl.def file, which is produced by the x_lapw script.
Please edit $WIENROOT/x_lapw.
Goto "case qtl:" and edit or replace the whole qtl-block by:
case qtl:
if ($?make_inq) then #Kevin Jorissen, for use with telnes3
write_inq_lapw
endif
#we need lapw2 -fermi mandatory
if( "$para" == "para" ) then
x lapw2 -p -fermi $soqtl $updnqtl $hfqtl
else
x lapw2 -fermi $soqtl $updnqtl $hfqtl
endif
set exe = $command$para
set dopara = 1
cat << theend > $def
4,'$file.inso', 'unknown', 'formatted',0
5,'$file.inq', 'old', 'formatted',-1
6,'$file.outputq$updn','unknown','formatted',0
8,'$file.scf2$hf$updn', 'old', 'formatted',-1
9,'${scratch}$file.vector$hf$so$updn', 'unknown','unformatted',0
10,'${scratch}$file.vector$hf$so$dnup', 'unknown','unformatted',0
12,'$file.dmat$updn', 'unknown','formatted',0
16,'$file.qtl$updn', 'unknown','formatted',0
18,'$file.vsp$updn', 'old', 'formatted',-1
19,'$file.vsp$dnup', 'unknown','formatted',0
20,'$file.struct', 'old', 'formatted',-1
21,'$file.scfq$updn', 'unknown', 'formatted',0
22,'$file.rotlm', 'unknown', 'formatted',0
26,'$file.weight$hf$updn', 'old', 'formatted',-1
29,'${scratch}$file.tmp$updn','SCRATCH','unformatted',9000
30,'$file.qtltext$updn','unknown','formatted',0
31,'$file.qtlatom$updn','unknown','formatted',0
59,'$file.energy$hf$so$dnup', 'unknown','formatted',0
60,'$file.energy$hf$so$updn', 'unknown','formatted',0
theend
###45,'$file.normsoup','unknown','formatted',0
###46,'$file.normsodn','unknown','formatted',0
set natom=`head -2 $file.inq | tail -1 | cut -c1-10`
set iatom=1
while ($iatom <= $natom)
set cf=`expr 31 + $iatom`
set int=`expr 70 + $iatom`
echo "$cf,'$file.cf$iatom','unknown','formatted',0">>$def
echo "$int,'$file.int$iatom$updn','unknown','formatted',0">>$def
@ iatom ++
end
if ($cmplx == c ) then
echo " 7,'$file.in1c', 'unknown', 'formatted',-1">>$def
endif
breaksw
In particular the "-1" in the $file.vectordn line is wrong and causes
your problem.
----------------------
Another problem may persist in the qtlpara_lapw script, although this is
for spin-polarized calculations.
Please goto https://github.com/gsabo/WIEN2k-Patches/tree/master/21.1
and download/apply the patch for qtlpara_lapw
----------------------------
PS: Please always try a search of the mailing list at www.wien2k.at
(reg_user; mailinglist). Often your problem has been found previously
and usually was also solved there.
PPS: Unfortunately the description of L23 spectra of TMOs (like anatase)
will not be great. You will never get the correct branching ratio in
conventional DFT (even with core holes) calculations. See
Physical Review B 82, 205105 (2010)
Best regards
Peter Blaha
Am 01.12.2022 um 06:13 schrieb Laurence Marks:
>
>
> ---
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> www.numis.northwestern.edu <http://www.numis.northwestern.edu>
> "Research is to see what everybody else has seen, and to think what
> nobody else has thought" Albert Szent-Györgyi
>
> ---------- Forwarded message ---------
> From: *Laurence Marks* <laurence.marks at gmail.com
> <mailto:laurence.marks at gmail.com>>
> Date: Thu, Dec 1, 2022, 09:58
> Subject: Re: [Wien] parallel case.vector and qtl
> To: Tim Williams <timothy.williams at monash.edu
> <mailto:timothy.williams at monash.edu>>
>
>
> gfortran --version please.
>
> There have been several reports where similar errors occur [1], which
> others have handled. I don't understand this one [2], so I am leaving it
> to Peter who, I think, sorted it out before.
>
> [1] Do a Google search.
> [2] I don't understand how "unformatted" can be an illegal form.
>
> --
> Professor Laurence Marks
> Department of Materials Science and Engineering, Northwestern University
> www.numis.northwestern.edu <http://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 Thu, Dec 1, 2022, 09:23 Tim Williams <timothy.williams at monash.edu
> <mailto:timothy.williams at monash.edu>> wrote:
>
> Hi Laurence,
>
> fortran: gfortran
> c:gcc
> parallel:
>
> and
>
> mcem-admin at MU0019202:~/wien2k$ cat WIEN2k_OPTIONS
> current:FOPT:-ffree-form -O2 -ftree-vectorize -march=native
> -ffree-line-length-none -ffpe-summary=none
> current:FPOPT:-ffree-form -O2 -ftree-vectorize -march=native
> -ffree-line-length-none -ffpe-summary=none -fallow-argument-mismatch
> current:OMP_SWITCH:-fopenmp
> current:OMP_SWITCHP:-fopenmp
> current:LDFLAGS:$(FOPT) -L../SRC_lib
> current:DPARALLEL:'-DParallel'
> current:R_LIBS:/home/lib/libopenblas.so
> /usr/lib/x86_64-linux-gnu/liblapack.so.3 -lpthread
> current:RP_LIBS:-lredist -ltools -lfblacs -lblacs -lmpi
> current:FFTWROOT:/home/mcem-admin/fftw-3.3.10/
> current:FFTW_VERSION:FFTW3
> current:FFTW_LIB:lib64
> current:FFTW_LIBNAME:fftw3
> current:LIBXCROOT:
> current:LIBXC_FORTRAN:
> current:LIBXC_LIBNAME:
> current:LIBXC_LIBDNAME:
> current:SCALAPACKROOT:
> current:SCALAPACK_LIBNAME:
> current:BLACSROOT:
> current:BLACS_LIBNAME:
> current:ELPAROOT:
> current:ELPA_VERSION:
> current:ELPA_LIB:
> current:ELPA_LIBNAME:
> current:MPIRUN:mpirun -np _NP_ -machinefile _HOSTS_ _EXEC_
> current:CORES_PER_NODE:1
> current:MKL_TARGET_ARCH:
> mcem-admin at MU0019202:~/wien2k$
>
> As I noted, I’m new to all this so there are probably stupid errors
> everywhere….
>
> Cheers!
>
> Tim.
>
> ---
> Dr. Tim Williams
> Transmission Electron Microscope Manager
>
> Monash Centre for Electron Microscopy (MCEM)
> Monash University
> Room 103, 10 Innovation Walk, Clayton Campus
> Wellington Road
> Clayton VIC 3800
> Australia
> T: +61 (0) 3 9902 0721
> M: +61 (0) 401 853 850
> e: timothy.williams at monash.edu <mailto:timothy.williams at monash.edu>
>
> CRICOS Provider: Monash University
> 00008C/01857J
>
> We acknowledge and pay respects to the Elders
> and Traditional Owners of the land on which our
> four Australian campuses stand.
>
> On 1 Dec 2022, at 14:39, Laurence Marks <laurence.marks at gmail.com
> <mailto:laurence.marks at gmail.com>> wrote:
>
> It was much too early in the morning when I responded.
>
> For whatever reason, your version appears to dislike the
> 'unformatted' file descriptor. What is your compiler (including
> version) and compile options.
>
> ---
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> www.numis.northwestern.edu <http://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 Thu, Dec 1, 2022, 07:15 Tim Williams <timothy.williams at monash.edu
> <mailto:timothy.williams at monash.edu>> wrote:
>
> Hi Laurence,
>
> only qtl.def, no others. Here’s the story…
>
> I’m pretty sure I’m not doing spin polarised calculations!
>
> Cheers,
>
> Tim.
>
>
> mcem-admin at MU0019202:~/wien2k/anatasenosym$ ls *qtl*.def
> qtl.def
> mcem-admin at MU0019202:~/wien2k/anatasenosym$ x qtl -p -d
> mcem-admin at MU0019202:~/wien2k/anatasenosym$ cat qtl.def
> 4,'anatasenosym.inso', 'unknown', 'formatted',-1
> 5,'anatasenosym.inq', 'old', 'formatted',-1
> 6,'anatasenosym.outputq','unknown','formatted',0
> 8,'anatasenosym.scf2', 'old', 'formatted',-1
> 9,'./anatasenosym.vector', 'unknown','unformatted',-1
> 10,'./anatasenosym.vectordn', 'unknown','unformatted',-1
> 12,'anatasenosym.dmat', 'unknown','formatted',0
> 16,'anatasenosym.qtl', 'unknown','formatted',0
> 18,'anatasenosym.vsp', 'old', 'formatted',-1
> 19,'anatasenosym.vspdn', 'unknown','formatted',-1
> 20,'anatasenosym.struct', 'old', 'formatted',-1
> 21,'anatasenosym.scfq', 'unknown', 'formatted',0
> 22,'anatasenosym.rotlm', 'unknown', 'formatted',0
> 26,'anatasenosym.weight', 'old', 'formatted',-1
> 29,'./anatasenosym.tmp','SCRATCH','unformatted',9000
> 30,'anatasenosym.qtltext','unknown','formatted',0
> 31,'anatasenosym.qtlatom','unknown','formatted',0
> 59,'anatasenosym.energydn', 'unknown','formatted',-1
> 60,'anatasenosym.energy', 'unknown','formatted',-1
> 32,'anatasenosym.cf1','unknown','formatted',0
> 71,'anatasenosym.int1','unknown','formatted',0
> 7,'anatasenosym.in1c', 'unknown', 'formatted',-1
> mcem-admin at MU0019202:~/wien2k/anatasenosym$ x qtl -p
> running QTL in parallel mode
> calculating QTL's from parallel vectors
> STOP QTL – Error
> 0.071u 0.012s 0:00.10 80.0%0+0k 616+32io 7pf+0w
> mcem-admin at MU0019202:~/wien2k/anatasenosym$ ls -lt
> anatasenosym.vector*
> -rw-rw-r-- 1 mcem-admin mcem-admin 56340806 Dec 1 12:30
> anatasenosym.vector_1
> -rw-rw-r-- 1 mcem-admin mcem-admin 53127908 Dec 1 12:30
> anatasenosym.vector_4
> -rw-rw-r-- 1 mcem-admin mcem-admin 55727818 Dec 1 12:30
> anatasenosym.vector_2
> -rw-rw-r-- 1 mcem-admin mcem-admin 54803134 Dec 1 12:30
> anatasenosym.vector_3
> -rw-rw-r-- 1 mcem-admin mcem-admin 0 Nov 30 14:21
> anatasenosym.vector
> mcem-admin at MU0019202:~/wien2k/anatasenosym$
>
> The .vector is from earlier serial run.
>
>
>
>
>
> ---
> Dr. Tim Williams
> Transmission Electron Microscope Manager
>
> Monash Centre for Electron Microscopy (MCEM)
> Monash University
> Room 103, 10 Innovation Walk, Clayton Campus
> Wellington Road
> Clayton VIC 3800
> Australia
> T: +61 (0) 3 9902 0721
> M: +61 (0) 401 853 850
> e: timothy.williams at monash.edu <mailto:timothy.williams at monash.edu>
>
> CRICOS Provider: Monash University
> 00008C/01857J
>
> We acknowledge and pay respects to the Elders
> and Traditional Owners of the land on which our
> four Australian campuses stand.
>
> On 1 Dec 2022, at 12:14, Laurence Marks
> <laurence.marks at gmail.com <mailto:laurence.marks at gmail.com>> wrote:
>
> My next guess is that somehow you have spin-polarized set in w2web.
>
> Or you somehow have "dn" set. First, do "ls *qtl*.def". You
> should only have qtl.def. if you have dnqtl.def (or upqtl.def)
> that seems odd.
>
> If you do try "rm *.def ; x qtl -p -d". You should now have a
> qtl.def file inside of which there are no "up" or "dn". (Thus
> does not run qtl, just sets it up.) If that is the case then x
> qtl -p should work.
>
> Beyond that I have no guesses at the moment.
>
> ---
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> www.numis.northwestern.edu <http://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 Thu, Dec 1, 2022, 06:32 Tim Williams
> <timothy.williams at monash.edu
> <mailto:timothy.williams at monash.edu>> wrote:
>
> Hi Laurence,
>
> Thanks for your reply, I’m currently looking at Ti-O in
> non-magnetic systems (eg just anatase, rutile etc). non-spin
> polarised. My interest is ELNES of the d electrons, L23
> edges. The non-parallel calculations work fine! The neutral
> structures work fine serially but I need to run parallel
> when putting in core holes with supercells.
>
> Eventually I’d like to look at Mn, Fe and again the L23.
>
> Best,
>
> Tim.
>
>
>
>
>
> ---
> Dr. Tim Williams
> Transmission Electron Microscope Manager
>
> Monash Centre for Electron Microscopy (MCEM)
> Monash University
> Room 103, 10 Innovation Walk, Clayton Campus
> Wellington Road
> Clayton VIC 3800
> Australia
> T: +61 (0) 3 9902 0721
> M: +61 (0) 401 853 850
> e: timothy.williams at monash.edu
> <mailto:timothy.williams at monash.edu>
>
> CRICOS Provider: Monash University
> 00008C/01857J
>
> We acknowledge and pay respects to the Elders
> and Traditional Owners of the land on which our
> four Australian campuses stand.
>
> On 1 Dec 2022, at 09:48, Laurence Marks
> <laurence.marks at gmail.com <mailto:laurence.marks at gmail.com>>
> wrote:
>
> It sounds like you are using "run_lapw" for your
> calculations. This is for non-spinpolarized cases, which is
> correct for many systems. If a material is magnetic, for
> instance Fe, then you have to use "runsp_lapw" which will
> produce up & dn spins. However, this is twice as slow.
>
> For special non-magnetic cases when you are using an orbital
> correction (e.g. for d-electrons) then runsp_c is useful.
>
> Which are you using?
>
> --
> Professor Laurence Marks
> Department of Materials Science and Engineering,
> Northwestern University
> www.numis.northwestern.edu <http://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 Thu, Dec 1, 2022, 02:54 Tim Williams via Wien
> <wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>> wrote:
>
> Greetings,
>
> I’m a very new Wien2k user and I am attempting to
> calculate DOS and ELNES for (mainly) transition metal
> oxides. I have installed Wien2k 21.1 on a single desktop
> with Intel i7. It’s slow but I hope to move to an HPC
> system. I am using both W2WEB and command line.
>
> Whilst I’m getting results in serial calculations
> (except that for some reason my calculations generate
> only case.vector and not case.vectordn, but I can cp to
> the correct name), parallel calculations are stalled
> after the SCF as qtl can’t read my separate
> case.vector_n files. The parallel scf seems to execute
> correctly without errors and generates separate
> case.vector_n files.
>
> All the case.vector_n files (n=1~4) are in the same
> local directory. Running x qtl -p gives the error as
> follows in qtl.error:
>
> ‘QTL’ - can’t open unit: 10
>
> ‘QTL’ - filename: ./case.vectordn
>
> ‘QTL’ - status: unknown form: unformatted
>
> In this directory /case/ are 4 vector files, i.e.
> case.vector_1~4
>
> I have found (in the archive) and applied a patch to
> qtlpara_lapw,
>
> https://github.com/gsabo/WIEN2k-Patches/tree/master/21.1
> <https://github.com/gsabo/WIEN2k-Patches/tree/master/21.1>
>
> But this had no effect. It appears that qtl is
> attempting to read only non-existant “case.vectordn”
> rather than the parallel case.vector_1~4 files.
>
> I’d appreciate any advice to help me get parallel
> calculations working. Beginner so please be gentle….
>
> Cheers,
>
> Tim.
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> <mailto: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>
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html <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
--
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300
Email: peter.blaha at tuwien.ac.at WIEN2k: http://www.wien2k.at
WWW: http://www.imc.tuwien.ac.at
-------------------------------------------------------------------------
More information about the Wien
mailing list