Re: [Wien] Fold2Bloch in parallel mode

[Rubel, Oleg]

The multiplicity 1:1:1 implies that the primitive cell and the supercell are 
the same. Thus, there is nothing to unfold. You should simply plot a classical 
band structure in this case.

Thank you
Oleg
--
Oleg Rubel (PhD, PEng)
Department of Materials Science and Engineering
McMaster University
JHE 359, 1280 Main Street West, Hamilton, Ontario L8S 4L8, Canada
Email: rub...@mcmaster.ca
Tel: +1-905-525-9140, ext. 24094
Web: http://olegrubel.mcmaster


From: Wien [wien-boun...@zeus.theochem.tuwien.ac.at] on behalf of Tristan de 
Boer [tristan.deb...@usask.ca]
Sent: October 18, 2017 17:31
To: wien@zeus.theochem.tuwien.ac.at
Subject: Re: [Wien] Fold2Bloch in parallel mode

Thanks for the replies. Using join_vectorfiles is working well. I have a
follow up question. I've been able to replicate the tutorials, and a
supercell of my own, but when I use fold2bloch on a single unit cell
with folds of 1:1:1 the weights in the case.f2b file are all 1. Do you
have any comment on why this might be?
Best Regards,
Tristan

On 2017-10-18 9:12 AM, Rubel, Oleg wrote:
> Either solution (Peter’s or Maciej’s) should work. The user’s choice might 
> depend on the size of the vector files.
>
> If you process individual *.vector_X files with fold2Bloch, it will warn that 
> the number of k-points is less than in the klist-file. This warning can be 
> ignored since k-points are split between *.vector_X files.
>
> Thanks
> Oleg
>
> --
> Oleg Rubel (PhD, PEng)
> Department of Materials Science and Engineering
> McMaster University
> JHE 359, 1280 Main Street West, Hamilton, Ontario L8S 4L8, Canada
> Email: rub...@mcmaster.ca
> Tel: +1-905-525-9140, ext. 24094
> Web: http://olegrubel.mcmaster.ca
>
>> On Oct 18, 2017, at 10:38, Peter Blaha  wrote:
>>
>> There is a   join_vectorfile   utility which combines the parallel vector 
>> files into a single one.
>>
>> On 10/17/2017 07:21 PM, Tristan de Boer wrote:
>>> Dear WIEN2k Users,
>>> I'm trying to use Fold2Bloch for a calculation in which I've done
>>> 'x lapw1 -c -p -band'. In this case, the case.vector file is empty, and
>>> I have several case.vector_# files. If I simply execute fold2bloch using
>>> the case.vector_1 as an input it warns me that the case.klist file has
>>> more K-points than the vector file. Is there a way to provide as input
>>> multiple case.vector_# files, or can I only use this program having done
>>> 'x lapw1 -c -band' in a non-parallel mode?
>>> Best Regards,
>>> Tristan
>>> ___
>>> 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/TC_Blaha
>> --
>> ___
>> 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
___
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] Fold2Bloch in parallel mode

[Tristan de Boer]

Thanks for the replies. Using join_vectorfiles is working well. I have a 
follow up question. I've been able to replicate the tutorials, and a 
supercell of my own, but when I use fold2bloch on a single unit cell 
with folds of 1:1:1 the weights in the case.f2b file are all 1. Do you 
have any comment on why this might be?

Best Regards,
Tristan

On 2017-10-18 9:12 AM, Rubel, Oleg wrote:

Either solution (Peter’s or Maciej’s) should work. The user’s choice might 
depend on the size of the vector files.

If you process individual *.vector_X files with fold2Bloch, it will warn that 
the number of k-points is less than in the klist-file. This warning can be 
ignored since k-points are split between *.vector_X files.

Thanks
Oleg

--
Oleg Rubel (PhD, PEng)
Department of Materials Science and Engineering
McMaster University
JHE 359, 1280 Main Street West, Hamilton, Ontario L8S 4L8, Canada
Email: rub...@mcmaster.ca
Tel: +1-905-525-9140, ext. 24094
Web: http://olegrubel.mcmaster.ca


On Oct 18, 2017, at 10:38, Peter Blaha  wrote:

There is a   join_vectorfile   utility which combines the parallel vector files 
into a single one.

On 10/17/2017 07:21 PM, Tristan de Boer wrote:

Dear WIEN2k Users,
I'm trying to use Fold2Bloch for a calculation in which I've done
'x lapw1 -c -p -band'. In this case, the case.vector file is empty, and
I have several case.vector_# files. If I simply execute fold2bloch using
the case.vector_1 as an input it warns me that the case.klist file has
more K-points than the vector file. Is there a way to provide as input
multiple case.vector_# files, or can I only use this program having done
'x lapw1 -c -band' in a non-parallel mode?
Best Regards,
Tristan
___
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/TC_Blaha
--
___
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

Re: [Wien] Fold2Bloch in parallel mode

[Rubel, Oleg]

Either solution (Peter’s or Maciej’s) should work. The user’s choice might 
depend on the size of the vector files.

If you process individual *.vector_X files with fold2Bloch, it will warn that 
the number of k-points is less than in the klist-file. This warning can be 
ignored since k-points are split between *.vector_X files.

Thanks
Oleg

--
Oleg Rubel (PhD, PEng)
Department of Materials Science and Engineering
McMaster University
JHE 359, 1280 Main Street West, Hamilton, Ontario L8S 4L8, Canada
Email: rub...@mcmaster.ca
Tel: +1-905-525-9140, ext. 24094
Web: http://olegrubel.mcmaster.ca

> On Oct 18, 2017, at 10:38, Peter Blaha  wrote:
> 
> There is a   join_vectorfile   utility which combines the parallel vector 
> files into a single one.
> 
> On 10/17/2017 07:21 PM, Tristan de Boer wrote:
>> Dear WIEN2k Users,
>> I'm trying to use Fold2Bloch for a calculation in which I've done
>> 'x lapw1 -c -p -band'. In this case, the case.vector file is empty, and
>> I have several case.vector_# files. If I simply execute fold2bloch using
>> the case.vector_1 as an input it warns me that the case.klist file has
>> more K-points than the vector file. Is there a way to provide as input
>> multiple case.vector_# files, or can I only use this program having done
>> 'x lapw1 -c -band' in a non-parallel mode?
>> Best Regards,
>> Tristan
>> ___
>> 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/TC_Blaha
> --
> ___
> 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] Fold2Bloch in parallel mode

[Peter Blaha]

There is a   join_vectorfile   utility which combines the parallel 
vector files into a single one.


On 10/17/2017 07:21 PM, Tristan de Boer wrote:

Dear WIEN2k Users,
I'm trying to use Fold2Bloch for a calculation in which I've done
'x lapw1 -c -p -band'. In this case, the case.vector file is empty, and
I have several case.vector_# files. If I simply execute fold2bloch using
the case.vector_1 as an input it warns me that the case.klist file has
more K-points than the vector file. Is there a way to provide as input
multiple case.vector_# files, or can I only use this program having done
'x lapw1 -c -band' in a non-parallel mode?
Best Regards,
Tristan
___
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/TC_Blaha
--
___
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] optimization of an orthorhombic cell

[Dr. K. C. Bhamu]

Thank you very much Prof. Tomas for letting me the nice tool.

I tried the OrthOpt script as usual as suggested "source volumeOrtho"  in a
pre-initialized case directory on a i3 Laptop with Wien2k_17.1.
It showed me several errors on screen (just copying few error messages
while the optimization process is still running):

--
 beforeCompass N says it has been  already computed
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
position NE is A = 33.0   B = 0.5
energy NE is

parent directory:
/home/bhamu/wien_work/palmexOrthOptA1.00
beforeCompass NE says it will be computed soon
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
position E is A = 33.0   B = 0.0
energy E is

parent directory:
/home/bhamu/wien_work/palmexOrthOptA1.00
beforeCompass E says it will be computed soon
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
position SE is A = 33.0   B = -0.5
energy SE is

parent directory:
/home/bhamu/wien_work/palmexOrthOptA1.00
beforeCompass SE says it will be computed soon
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
position S is A = 32.5   B = -0.5
energy S is

beforeCompass S says it has been  already computed
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory

>   stop error
cleaning /home/bhamu/wien_work/palmexOrthOptA-1.00/arielOrthOptSE/OrthOpt
position SW is A = 32.0   B = -0.5
energy SW is

beforeCompass SW says it has been  already computed
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
position W is A = 32.0   B = 0.0
energy W is

beforeCompass W says it has been  already computed
rm: cannot remove '/home/bhamu/wien_work/palmexOrthOptA1.00/arielOrthOpt*':
No such file or directory
in /home/bhamu/wien_work/palmexOrthOptA-1.00/arielOrthOptSE/OrthOpt
position NW is A = 32.0   B = 0.5
energy in SE :

>   stop error






>   stop error
cleaning /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptE/OrthOpt
in /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptE/OrthOpt
energy in E :

>   stop error
cleaning /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptSE/OrthOpt
in /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptSE/OrthOpt
energy in SE :

>   stop error
cleaning /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptNE/OrthOpt
in /home/bhamu/wien_work/palmexOrthOptA0.00/arielOrthOptNE/OrthOpt
energy in NE :

--


In addition to above, it is mentioned in "compute.job"  that the script
will use 15 core by default. with below statements:

 # If you want some more parallelization then mind that the number of
 # processors increases as 15, 30, 45, 60 and so on
#cat > .machines <<+
#1:${HOSTNAME}:1
#+ # The + sign must be at the very beginning of the line



So does it mean that we do not need to modify anything if we want to use
only 15 cores in main cluster job file provided that the core must be
greater then 15 in cluster job file?

Suppose in my cluster job file I specify 16 cores and then the "source
volumeOrtho" should use 15 cores. Is it?

What about if we use 30 cores? how to change above lines in "compute.job"
to use 30 cores?


Kind regards
Bhamu




On Wed, Oct 18, 2017 at 2:11 PM, Tomas Kana  wrote:

> Dear Bhamu,
> There is an alternative way. You may use the
> OrthoOpt package available at
> http://susi.theochem.tuwien.ac.at/reg_user/unsupported/OrthoOpt/
> It proceeds in a similar way Stefaan has suggested in previous
> mail copied below. It first fixes the volume and finds
> a minimum energy depending on two variables: c/a and b/a.
> Then, it moves to the next volume and so on until it reaches
> minimum volume.
> HTH
> Tomas
>
> No, not really. You have three (unit cell) degrees of freedom: E(a,b,c).
> Or similarly E(V, b/a, c/a)   (with V=a*b*c). The different options in
> optimize_lapw give you a series of unit cells with either different values
> of V (and b/a and c/a constant), different values of b/a (and V and c/a
> constant) and different values of c/a (and V and b/a constant). These are
> three slices through the energy landscape. By finding the energy minimum
> along one slice, fix it, then go on with the next slice, etc. … you can
> gradually approach the absolute energy minimum. It might take several
> iterations through this scheme if your starting cell is far away from the
> minimum. Best is to start first with the degree of freedom that affects
> energy most (usually the volume).
>
>
>
> Stefaan
>
> Předmět: Re: [Wien] optimization of an orthorhombic cell
> Dear List,
> Could you please help me to get optimize lattice
> parameters from the procedure that I followed as
> mentioned below:
> The procedure is a four-step task what I am following:
> The original 

Re: [Wien] optimization of an orthorhombic cell

[Tomas Kana]

Dear Bhamu,
There is an alternative way. You may use the 
OrthoOpt package available at 
http://susi.theochem.tuwien.ac.at/reg_user/unsupported/OrthoOpt/
It proceeds in a similar way Stefaan has suggested in previous 
mail copied below. It first fixes the volume and finds 
a minimum energy depending on two variables: c/a and b/a. 
Then, it moves to the next volume and so on until it reaches
minimum volume. 
HTH 
Tomas 

No, not really. You have three (unit cell) degrees of freedom: E(a,b,c). Or 
similarly E(V, b/a, c/a)   (with V=a*b*c). The different options in 
optimize_lapw give you a series of unit cells with either different values of V 
(and b/a and c/a constant), different values of b/a (and V and c/a constant) 
and different values of c/a (and V and b/a constant). These are three slices 
through the energy landscape. By finding the energy minimum along one slice, 
fix it, then go on with the next slice, etc. … you can gradually approach the 
absolute energy minimum. It might take several iterations through this scheme 
if your starting cell is far away from the minimum. Best is to start first with 
the degree of freedom that affects energy most (usually the volume).

 

Stefaan

Předmět: Re: [Wien] optimization of an orthorhombic cell
Dear List,
Could you please help me to get optimize lattice 
parameters from the procedure that I followed as 
mentioned below:
The procedure is a four-step task what I am following:
The original lattice parameters were: a, b, c=16.763005, 24.803600, 16.320431   
  vol 6785.64777697697 bohr^3.
1.   VARY C/A RATIO with CONSTANT VOLUME and B/A (orthorh lattice)
getting below optimized data: 
 ==
Lowest point on X-axis is =  -1.615800
Minimum value of B/A is = 1.455755
Minimum value of ENE is =   -339059.104498 Ry
 ==
Value of A is = 16.85428 bohr  ;8.91890 Ang
Value of B is = 24.53569 bohr  ;   12.98373 Ang
Value of C is = 16.40929 bohr  ;8.68342 Ang
 ==
Initial value of  B/A is =  1.479663
Conv. Unit Cell Vol =  6785.7557 bohr^3 ; **Ang^
2.  Volume optimization using option 1 in "x optimize" and then calculating new 
volume from eplot option.
Query: Step two is giving optimized volume but I do not know how to calculate 
lattices parameters from a single volume
value for the next step "VARY C/A RATIO with CONSTANT VOLUME and B/A (orthorh 
lattice)"?
The original cell volume and lattice parameters are mentioned at the beginning.
and the new volume from step two is: 6815.3427 bohr^3
The above same query is for the remaining process:
3. optimization of C/A 
4. optimization of V.
*5 At the final step I will do structural 
relaxation (min_.).
Could you please advice me how to proceed?
Kind regards,
Bhamu
___
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] Optimization of monoclinic structure

[Tomas Kana]

Thanks Gavin for reply.
 I just add that OrthoOpt fixes  the volume and 
relaxes simultaneously c/a and b/a ratio of lattice 
parameters using several times x optimize and searching 
the best combination of c/a and b/a. 
Then it moves on for next volume and so on.
It does not change the angles of the lattice.
Tomas 

Předmět: Re: [Wien] Optimization of monoclinic structure

I don't remember.  Probably all parameters are fixed except for one of 
them, then this is repeated separately for each parameter:

A, B, C fixed with changing GAMMA value, then
A, B, GAMMA fixed with changing C value, then
...
B, C, GAMMA fixed with changing A value

"x optimize" or OrthoOpt probably automatically set up the set of 
calculations in this way in a job script like optimize.job and then the 
job script is used to run all the calculations [ 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg16418.html 
].

 From the sounds of it, you are setting up each calculation by hand 
instead of using one of the above scripts/programs to help with that.

For monoclinic optimization, make sure the monoclinic angle (i.e., the 
one angle that is not equal to 90) is set as gamma.  As maybe this could 
cause the "SELECT" error due a symmetry break [ 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg12124.html 
].

The "SELECT" error could also be caused by other things.  See the "NO 
ENERGY LIMITS FOUND IN SELECT" on page 231 under section "12 Trouble 
shooting" in the WIEN2k 17.1 usersguide [ 
http://susi.theochem.tuwien.ac.at/reg_user/textbooks/usersguide.pdf ], 
where is says:

"Usually this happens when your input is not ok, or for very ill 
conditioned problems (very rare), or more likely, when “Ghostbands” 
appeared (or some states were missing) because of bad energy parameters 
in case.in1."

input is not ok =  In an optimization, you are distorting the structure 
(changing the lattice parameters).  It is probably straightforward that 
if you distort a structure too much (% change is too large)  while you 
are doing this, you could possible generate a bad case.struct file.

As the above sentence also says, it could be caused by ghost bands. To 
help with that, there is section "12.1 Ghost bands" starting on page 232 
in the WIEN2k usersguide.

A related error might be the QTL-B error.  Refer to the links found in 
the post at:

https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg12625.html

On 10/15/2017 8:26 PM, Shalika R. Bhandari wrote:
  Dear all, Thank you for your suggestions .
   Now I have a bit confusion like while optimizing monoclinic structure 
  does A, B,C,and GAMMA changes at the same time or we need to do them 
  separately?
  Also while running optimization, after first cycle it shows ERROR and 
  stops running. LIKE THIS..
  46.9u 0.0s 0:47.75 98.4% 0+0k 3696+40656io 12pf+0w
  1.3u 0.0s 0:01.59 89.3% 0+0k 73552+35336io 6pf+0w
  clmextrapol_lapw has generated a new RELAX.clmdn
  hup: Command not found.
  changing RELAX.in2c
  changing RELAX.in2_ls
  changing RELAX.in2_st
  changing RELAX.in2_sy
   LAPW0 END
  SELECT - Error
 
  stop error
  ERROR status in RELAX_mon1.00
 
  I Want some more suggestions to resolve this problem sir.
___
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