[Wien] Fw: Optical property calculation when spin orbit added

[Gavin Abo]

Regarding the empty case.vectorsoup, probably it was created when lapwso 
started, but it likely failed while running before any data could be 
written to it.  However, I doubt anyone can be of further help.  The 
case.dayfile, standard input/output file(s), *.error, and/or other 
output files are typically needed to locate the cause.

My guess is that you submit your job script in the terminal with [1,2]:

qsub pbs.job

Unless you are using a script for the w2web interface [3].

For example, pbs.job at

http://susi.theochem.tuwien.ac.at/reg_user/faq/pbs.html

has in it:

#define here your WIEN2k command

run_lapw -p -i 40 -fc 1. -I

Yes, you should be able to run more the one WIEN2k command in it.  I 
don't have a system with qsub to test it, but for example, changing 
those lines to the following might work:

#define here your WIEN2k command

init_lapw -b -sp
initso_lapw < initso_input_file
runsp_lapw -so -orb -ec 0.0001 -cc 0.001
x kgen -so < kgenso_input_file
x lapw1 -up
x lapw1 -dn
x lapwso -up
x lapw2 -so -fermi -up
x lapw2 -so -fermi -dn
x lcore -up
x lcore -dn
# Can keep adding additional commands here

As seen above, you may need to use "<" to get input from a file instead 
of the keyboard [4,5].

In addition, the optic steps in your email appear to be for a serial 
calculation.  However, the "-p" in your runsp_lapw indicates you did a 
parallel calculation for the scf, which may be why your calculation does 
not work.

I think you need to check if a program needs the -p option by running:

x -h program

or check the WIEN2k usersguide (UG).

For example, in the WIEN2k 17.1 UG [6] in section 8.11.1 on page 165, it 
shows that kram does not take a -p option.

I cannot recall for sure, but it might be that the optic programs were 
not parallelized.  I think join_vectorfiles might be needed because of 
that.  Refer to the WIEN2k UG (page 207 section 9.17).

[1] http://www.hpc.lsu.edu/docs/guides/software.php?software=wien2k
[2] https://docs.loni.org/wiki/WIEN2K_on_LONI_x86_Clusters
[3] 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg01658.html
[4] 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg14196.html
[5] http://linuxcommand.org/lc3_lts0070.php
[6] http://susi.theochem.tuwien.ac.at/reg_user/textbooks/usersguide.pdf

On 7/13/2017 11:25 AM, shaymlal dayananda wrote:
> Dear Dr. Abo
>
> Thank you for your advises.
> If I summarize the problem so far; I was trying to finish OPTIC for a 
> system with sp,Hubbard-U and SO
>
> 01. I had finished SCF with sp/U/SO. and continued OPTIC. When I run 
> "opticc -so -orb -up" , I got case.vectorsoup error
> When I first get case.vectorsoup error in 
> [http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg16062.html 
> ] , I didn't have set anything  for "export SCRACH" in my bashrc or in 
> job scripts.
> With this the error was
> ERROR:
> 'OPTIC' -  can't open unit: 10
>   'OPTIC' -  filename: /scratch/10820461.yak.local/SCFsp-U-SO.vectorsoup
>   'OPTIC' -  status: OLD          form: UNFORMATTED
> 02
> After your advice to set export SCRATCH=./ , we added export 
> SCRACH=$PBS_O_WORKDIR in job scripts instead of putting it bashrc. And 
> I tried a test case.
> Now I have empty case.vectorsoup and case.vectorsodn files after 
> completion of SCF with Hubbard-U and SO. With this I have 
> /global/scratch/WIEN2k/NiO-14/NiO-14.vectorsoup
>
> With this I didn't continue with OPTIC because the files are empty and 
> it will end up with error. I wanted to make it correct before checking 
> for OPTIC
> I checked xxxxxx.yak.local . It always change, when I submit jobs. 
> (ex. If I finish SCF and then submit for SO, it changes)
>
> Thus, as you said, may be it is better to change to 17.1 version. 
> Actually I feel tired going back and forth with all these troubles I 
> had up to now.
>
> However, I would like to know if I want to try everything is a single 
> job, Can I run all the commands below in a single script without going 
> through interface. ( I mean just submitting it through terminal)
>
> runsp_lapw -p -so  -dm  -orb  -ec 0.0001 -cc 0.001
> x kgen -so
> x lapw1 -up
> x lapw1 -dn
> x lapwso -up
> x lapw2 -so -fermi -up
> x lapw2 -so -fermi -dn
> x lcore -up
> x lcore -dnx lapw1 -up
> etc
>
> Thank you very much
> Chami
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20170713/cdfee472/attachment.html>