[Wien] question on optical property calculation

[shaymlal dayananda]

Dear Wien2k users
I am little bit confused with the definition of the calculated optical property 
components. For instant, Im_epsilon_xx; what is the correct definition of this?
Is this the "IM_epsilon along x-direction when the external electric field is 
parallel to x-direction?" 
Further, if I have added spin orbital coupling along [010] direction, does that 
definition to the Im_epsilon_xx change?
I have another question regarding the SCISSOR_OPERATOR. When I think, this is 
used to match the band gaps when we have a smaller theoretical value than the 
experimental one. If anyone have some idea, please let me know whether it is a 
must for any semiconductor (metal oxide)?
For my case my calculated band gap is greater than the experimental one. Is 
there a way to get this matched to the experimental one ?
I am very much grateful for your help on the success of my project.
Thank you in advanceChami
___
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] Problem of permission to submit the program and que setting

[Gavin Abo]

Since I don't have a system with Rocks and qsub, I cannot be of much 
help.  That looks like it might be a Grid Engine error and not a WIEN2k 
error, so you might get better help resolving that error on a Grid 
Engine mailing list [1] or a Rocks mailing list [2].


It looks like qsub doesn't allow the 'root' account to submit jobs [3].  
Did you run qsub under as root?  If so, that may be the cause of the 
error. The solution may be to run instead qsub under a 'user' account.


[1] http://gridengine.org/mailman/listinfo/users
[2] https://lists.sdsc.edu/mailman/listinfo/npaci-rocks-discussion
[3] 
https://stackoverflow.com/questions/37733095/unable-to-run-jobs-on-cfncluster


On 9/19/2017 4:04 AM, sharma.krishnaswaroop wrote:


Dear Prof. Blaha/WIEN users
Sorry incomplete message was sent by mistake. We are trying to set up 
a small cluster for WIEN2K , by using one server and 4 nodes, all i5 
intel desktops. We have successfully installed Rocks 6.2 and WIEN2K 
16.1. We are finding problem in que setting. We are using SGE for 
queing and have defined through qmon the administrative host, submit 
host and execution hosts. But when we submit a job by using qsub 
command it give error with the message "Unable to run job: job 
rejected: root@cluster.local is not allowed to submit jobs. Exiting." 
We tried to find out a solution from User guide and Frequently asked 
questions, but could not get a solution. Could any body help us to 
solve the problem and successfully instal the cluster.

With thanks in anticipation.
Dr. K. S. Sharma
The IIS UNIVERSITY JAIPUR INDIA

___
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] bash script to run initso_lapw for different M-axis

[Gavin Abo]

I don't know everything about SO.  You can read more about the 
EFG-MATRIX is a NULLMATRIX error in the posts at:


https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg03446.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg04411.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg06048.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg09677.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg05035.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg04409.html

This error is in WIEN2k 13.1, right? Does it happen in WIEN2k 17.1? On 
the updates page [ http://susi.theochem.tuwien.ac.at/reg_user/updates/ 
], you can see changes such as to init_so_lapw, make_inso_lapw, 
SRC_lapwso, and SRC_symmetso since WIEN2k 13.1.  It may be that 17.1 
already has a fix that will resolve this error.


The Error in Parallel LAPW2 might just be a result of the lapw0 error.

On 9/19/2017 8:40 PM, venkatesh chandragiri wrote:

Dear Prof. Victor and Prof. Gavin,

thanks for your reply. The script sent by Prof. Gavin works fine. I 
just kept , editor ="noeditor"   in my .bashrc file. Now the editor is 
not opening. However, while running this script there was error as 
given below (although it is not connected with the error in the script 
write-up).


==

>   lapw2 -up -p   -c -so   (10:13:36) running LAPW2 in parallel mode
**  LAPW2 crashed!
0.232u 0.242s 0:00.54 87.0% 0+0k 0+392io 0pf+0w
error: command   /public/home/venkatesh/WIEN2k/lapw2cpara -up -c -so 
uplapw2.def   failed

>   stop error



uplapw2.error

=

'FERMI5' -  emax too low in vector-file
**  testerror: Error in Parallel LAPW2

=

The case.dayfile at the earlier shows

==

>   lapw0 -p    (10:12:50) starting parallel lapw0 at Wed Sep 20 
10:12:50 CST 2017

 .machine0 : processors
running lapw0 in single mode
 WARNING: The EFG-MATRIX is a NULLMATRIX !
 WARNING: The EFG-MATRIX is a NULLMATRIX !

=

I am using k-point parallization  to run soc calcualtions

The SOC calculations for single M-axis works well through regular 
procedure (200 axis). However, when I used the script to run for i3 
{200,201,202,103}, the above error appears for all cases.


Kindly suggest me the possible reason for this error

thanking you

with best regards,

venkatesh
___
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] bash script to run initso_lapw for different M-axis

[venkatesh chandragiri]

Dear Prof. Victor and Prof. Gavin,

thanks for your reply. The script sent by Prof. Gavin works fine. I
just kept , editor ="noeditor"   in my .bashrc file. Now the editor is not
opening. However, while running this script there was error as given below
(although it is not connected with the error in the script write-up).

==

>   lapw2 -up -p   -c -so   (10:13:36) running LAPW2 in parallel mode
**  LAPW2 crashed!
0.232u 0.242s 0:00.54 87.0% 0+0k 0+392io 0pf+0w
error: command   /public/home/venkatesh/WIEN2k/lapw2cpara -up -c -so
uplapw2.def   failed
>   stop error



uplapw2.error

=

'FERMI5' -  emax too low in vector-file
**  testerror: Error in Parallel LAPW2

=

The case.dayfile at the earlier shows

==

>   lapw0 -p(10:12:50) starting parallel lapw0 at Wed Sep 20 10:12:50
CST 2017
 .machine0 : processors
running lapw0 in single mode
 WARNING: The EFG-MATRIX is a NULLMATRIX !
 WARNING: The EFG-MATRIX is a NULLMATRIX !

=

I am using k-point parallization  to run soc calcualtions

The SOC calculations for single M-axis works well through regular procedure
(200 axis). However, when I used the script to run for i3
{200,201,202,103}, the above error appears for all cases.

Kindly suggest me the possible reason for this error

thanking you

with best regards,

venkatesh
___
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] bash script to run initso_lapw for different M-axis

[Víctor Luaña Cabal]

On Tue, Sep 19, 2017 at 07:57:36AM +0800, venkatesh chandragiri wrote:
> Dear Prof. Gavin,
> 
> Thank you again for the script to create the input for initso_lapw  and it
> works fine now. However, the initso_lapw is asking to edit the following
> files.
> 
> 1. case.inso
> 
> 2. case.in1
> 
> 3. case.outsymso
> 
> 4. case.klist
> 
> Hence, kindly let me know (paste the script lines) to avoid the editing of
> those above files and continue the initso_lapw in background itself.

If you are going to iniciate running a program or script ending thw task with & 
sends it to the background and you have another task at hand to continue 
editing, for instance. So

$ script -options file &

or

run_code -options input_file output_file &

or

run_code -options < input_file > output_file &

will send the run_code to the background.

If the script or program is already running you can use crtl-Z to
interrupt it and gain a new  shell (cshell or bash shell, for instance)
Using bg will send the interrumped process to background and fg to
the fore ground.  > > thank you and looking forward to your reply.
> > with best regards, > > venkatesh


Best regards,
 Víctor Luaña
--
.  ."Never let your sense of morals prevent you from doing
   / `' \ what is right." 
  /(o)(o)\  -- Salvor Hardin, "Foundation"
 /`. \/ .'\  
/   '`'`   \ 
|  \'`'`/  | 
|  |'`'`|  | 
 \/`'`'`'\/
==(((==)))===+===
! Dr.Víctor Luaña, in silico chemist & prof. !
! Departamento de Química Física y Analítica !
! Universidad de Oviedo, 33006-Oviedo, Spain !
! e-mail:    !
! phone: +34-985-103491  fax: +34-985-103125 !
++
 GroupPage: 
 Articles:  
 git-hub:   
 ORCID: -0003-4585-4627; RID: H-2045-2015
___
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] Problem of permission to submit the program and que setting

[sharma.krishnaswaroop]

Dear Prof. Blaha/WIEN usersSorry incomplete message was sent by mistake. We are 
trying to set up a small cluster for WIEN2K , by using one server and 4 nodes, 
all i5 intel desktops. We have successfully installed Rocks 6.2 and WIEN2K 
16.1. We are finding problem in que setting. We are using SGE for queing and 
have defined through qmon the administrative host, submit host and execution 
hosts. But when we submit a job by using qsub command it give error with the 
message "Unable to run job: job rejected: root@cluster.local is not allowed to 
submit jobs. Exiting." We tried to find out a solution from User guide and 
Frequently asked questions, but could not get a solution. Could any body help 
us to solve the problem and successfully instal the cluster.With thanks in 
anticipation.Dr. K. S. SharmaThe IIS UNIVERSITY JAIPUR INDIA

Sent from my Samsung Galaxy smartphone.___
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] Problem of permission to run a job and que setting

[sharma.krishnaswaroop]

Dear Prof. Blaha/WIEN usersWe are trying to set up a small cluster for WIEN2K 
on Rocks, by using one server and 4 nodes, all i5 desktops 


Sent from my Samsung Galaxy smartphone.___
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] relaxation

[Tomas Kana]

Dear Shalika, 

Alternatively, you may use the package OrthoOpt 
available at 
http://susi.theochem.tuwien.ac.at/reg_user/unsupported/OrthoOpt/
This package automatically determines the orthorhombic or monoclinic 
lattice parameters a,b,c (the monoclinic angle is fixed). 
You can try this for several values of the monoclinic angle 
and optimize the angle in this way. The a,b,c will be optimized 
automatically, too.
Tomas 

See option "[7] VARY A, B, C and Gamma (4D-case) (monoclinic lattice)" 
in section "5.3.1 Lattice parameters (Volume, c/a, lattice parameters)" 
on page 74 of the WIEN2k 17.1 usersguide [ 
http://susi.theochem.tuwien.ac.at/reg_user/textbooks/usersguide.pdf ].

Since there are so many degrees of freedom (is a calculation for 4 
dimensions), the calculation is likely very computational expensive.  
Hopefully you have a cluster with many cores.

That should be just lattice parameter optimization.  If atomic position 
optimization is need as well, that may require even more computation.  
The "Structure optimization-notes (pdf)" on the textbook page [ 
http://susi.theochem.tuwien.ac.at/reg_user/textbooks/ ], I think 
describes simultaneous optimization of lattice parameters and atomic 
positions with a simpler case as an example pretty well.

On 9/18/2017 5:49 AM, Shalika R. Bhandari wrote:
  Hi sir,
  I want to learn the steps  optimisation of monoclinic crystal ?

___
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] multiple fatbands plot

[Gavin Abo]

I forgot, you may be right that WIEN2k can only plot one band character 
at a time.  To plot more than one, you might need to create your own 
plot using a graphing software (like Origin, gnuplot, matlab, etc.) or 
use Spaghetti-primavera.  See the unsupported page for Spaghetti-primavera:



http://susi.theochem.tuwien.ac.at/reg_user/unsupported/


On 9/19/2017 1:14 AM, Wen Fong Goh wrote:


11 and 12 correspond to dxz and dyz (not quite t2g in cubic sym.). 
Ideally, I would like to have a fatband for dxz and another (different 
color) for dyz. What I found is even though line 11 (jatom jtype ...) 
is repeated, only the last line will be plotted, it can easily be 
checked by setting the size on the first line to be large.


2 11    2.0    # jatom, jtype, size  of heavier 
plotting
2 12    0.2    # jatom, jtype, size  of heavier 
plotting



For the TiC example in the WIEN2k usersguide, jtype of 6 should 
correspond to t2g. What is 11 and 12 in your case?  Not sure if 11 or 
12 are too high of a number and don't correspond to anything.


The "line switch" might need to be set to 2 in your case.insp and 
sometimes I think the circle size (size of heavier plotting) of 0.2 
might be too small to show the fat bands.  You might have to increase 
it so that it doesn't hide itself as a thin line.

On 9/18/2017 9:30 PM, Wen Fong Goh wrote:

Hi, how do you plot multiple fatbands? I tried adding another line
in case.insp after jatom jtype ... line but it doesn't work. 2
11    0.2    # jatom, jtype, size  of heavier
plotting 2 12    0.2    # jatom, jtype, size 
of heavier plotting 


___
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 -- A Mailing list for WIEN2k users 


zeus.theochem.tuwien.ac.at
Your email address: Your name (optional): You may enter a privacy 
password below. This provides only mild security, but should prevent 
others from messing with ...


wien - The Mail Archive 


www.mail-archive.com
Messages by Thread [Wien] problem of energy gap within mBJ halim said. 
Re: [Wien] problem of energy gap within mBJ tran; Re: [Wien] problem 
of energy gap within mBJ ...





___
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

Re: [Wien] multiple fatbands plot

[Wen Fong Goh]

11 and 12 correspond to dxz and dyz (not quite t2g in cubic sym.). Ideally, I 
would like to have a fatband for dxz and another (different color) for dyz. 
What I found is even though line 11 (jatom jtype ...) is repeated, only the 
last line will be plotted, it can easily be checked by setting the size on the 
first line to be large.

2 112.0# jatom, jtype, size  of heavier plotting
2 120.2# jatom, jtype, size  of heavier plotting


For the TiC example in the WIEN2k usersguide, jtype of 6 should correspond to 
t2g. What is 11 and 12 in your case?  Not sure if 11 or 12 are too high of a 
number and don't correspond to anything.

The "line switch" might need to be set to 2 in your case.insp and sometimes I 
think the circle size (size of heavier plotting) of 0.2 might be too small to 
show the fat bands.  You might have to increase it so that it doesn't hide 
itself as a thin line.

On 9/18/2017 9:30 PM, Wen Fong Goh wrote:


Hi, how do you plot multiple fatbands? I tried adding another line in case.insp 
after jatom jtype ... line but it doesn't work. 2 110.2 
   # jatom, jtype, size  of heavier plotting 2 120.2
# jatom, jtype, size  of heavier plotting

___
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 -- A Mailing list for WIEN2k 
users
zeus.theochem.tuwien.ac.at
Your email address: Your name (optional): You may enter a privacy password 
below. This provides only mild security, but should prevent others from messing 
with ...

wien - The Mail 
Archive
www.mail-archive.com
Messages by Thread [Wien] problem of energy gap within mBJ halim said. Re: 
[Wien] problem of energy gap within mBJ tran; Re: [Wien] problem of energy gap 
within mBJ ...


___
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] Hardware (i7-7700k?) for WIEN2k

[Peter Blaha]

Without knowing the new i7-7700K myself, a few recommendations:

You do NOT need more memory than 64 GB (in fact we have at most 32 GB on 
our  workstations). If a calculation requires that much memory, the 
single processor speed is too slow anyway.


I'd go for the I7, because it has faster memory and faster clock speed.

The advantage of a Xeon is usually that that you can combine 2-4 CPUs on 
a single motherboard. But usually have pay a lot for this possibility.


PS: I'd certainly also get in addition a regular 2-4TB harddisk (costs 
nothing, but is very useful)


Peter Blaha

On 09/19/2017 06:20 AM, Gavin Abo wrote:

Some additional comments:

If you need an idea of about how computational intensive a WIEN2k
calculation might be for around 60 atoms per cell or more.  The links
below and links in those posts might be helpful:

https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg14035.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg01420.html
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg05784.html

The hardware specifications for the i7-7700K and E5-2623 v3 processors
should be at the links:

https://ark.intel.com/products/97129/Intel-Core-i7-7700K-Processor-8M-Cache-up-to-4_50-GHz
http://ark.intel.com/products/83354/Intel-Xeon-Processor-E5-2623-v3-10M-Cache-3_00-GHz

The specifications look fairly similar.  So maybe there would not be too
much of a difference in the WIEN2k calculation run times between the two.

The i7-7700K supports a Max Memory Size of 64 GB.  The E5-2623 v3 could
have a major advantage when it comes to supporting a Max Memory Size of
768 GB.  However, if the motherboard the E5-2623 v3 goes on does not
have memory expansion slots for more than 64 GB or if you never plan to
add additional memory modules to increase the memory beyond 64 GB, then
that would not matter.

The i7-7700K does have about a 1 GHz faster processor frequency and it
looks like it supports a faster RAM (DDR4-2133/2400, while the E5-2623
v3 supports a slower DDR4 with frequency of 1600 or 1866 MHz).  So this
might give it better performance than the i7-7700K if the calculation
uses less than 64 GB of RAM.  Above 64 GB, the workstation would likely
use virtual memory and disk caching may significantly slow a calculation
(whereas the E5-2623 v3 with more than 64 GB should extend the limit of
this RAM bottleneck).  The i7-7700K also supports DDR3L-1333/1600.  If
both the i7-7700K and E5-2623 v3 workstations happen to use the same
DDR3 1600 RAM, then no speed up or slow down is expected from the RAM
frequency.

More importantly than all that may be the launch date for the E5-2623 v3
of Q3'14, while Q1'17 for the i7-7700K.  The Xeon E5-2623 v3 has been
around awhile.  So Linux distributions most likely have drivers that
support this processor.  With the i7-7700K being so new, you might have
to be more cautious.  If you decide to get the i7-7700K, I recommend
checking that the Linux distribution, compiler (in particular if you
plan to use a non-Intel compiler like gfortran), and libraries (such as
a blas library with the non-Intel compiler) that you will be using are
able to support and recognize the processor.

As an example, I think it was the Intel HD Graphics 530 onboard the
Intel Skylake processors when they were first launched that didn't have
a good Linux driver [
https://askubuntu.com/questions/698168/cant-get-intel-hd-graphics-530-skylake-i7-6700-to-work
]. If I recall correctly, the graphics were broken (or of poor quality)
for several months until the drivers were finally released.

Of note, there is also seems to be a E5-2623 v4:

http://ark.intel.com/products/92980/Intel-Xeon-Processor-E5-2623-v4-10M-Cache-2_60-GHz

On 9/18/2017 6:54 PM, Yundi Quan wrote:

hi,

4 cores with 8 threads is probably OK for using WIEN2k to study
transition metal oxides. For post-processing tools, such as
wien2wannier, it requires more memory. But 64 GB is enough in most
cases. I once bought a Dell XPS with 4 cores and 8 threads, 16 GB
memory back in 2011. It worked well for most of my calculations. Hope
this helps.




On Mon, Sep 18, 2017 at 2:44 PM, Karsten Küpper > wrote:

Dear WIEN2k-community,

We want to buy a workstation dedicated to run WIEN2k.
Our aim is to calculate mostly transition metal oxides with unit
cells ranging from 24 - 128 atoms supercells.
We are thinking about a workstation with at least 4 cores, at
least 64 GB RAM, and a 1TB SSD as a starting point.

1) May that be a reasonable choice?

2) Has anybody experiences with the i7-7700K 4.2 GHz (maybe also
compared to Intel-Xeon E5-2623 processors),  as there is no
benchmark test available on the WIEN2k homepage by now.

Thanks for your efforts in advance!

Kind regards
Karsten Küpper




___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at