[Wien] Problem with sbatch job

[Neeraj Kulhari]

Dear Team,
I am working with WIEN2k on HPC. While I am running commands on the
terminal it is working fine. But when I am submitting the job script, I am
getting pbe.scf but it just contains commands used for running lapw. I am
pasting my job script.
I am stuck with that.
Thanks and regards
#!/bin/bash
#SBATCH --time=00:30:00# Walltime in hh:mm:ss! keep small time
#SBATCH --nodes=1  # Number of nodes   ! Keep it 4 but we can
change as per job conditions
#SBATCH --ntasks-per-node=24
#SBATCH --cpus-per-task=2  # Number of OpenMP threads for each MPI
process/rank  ! Do not change
#SBATCH --mem=30   # Per nodes memory request (MB),  max mem is
375GB  ! Do not change
#SBATCH --account=*  # no change
#SBATCH --job-name=relax   # change it as per case
#SBATCH --partition=g100_usr_prod# do not change
#module load wien2k
export OMP_NUM_THREADS=4

case=pbe
# Use , as list seperator
IFS=','
# Convert string to array
hcpus=($SLURM_JOB_CPUS_PER_NODE)
unset IFS

declare -a conv

# Expand compressed slurm array
for cpu in ${hcpus[@]}; do
 if [[ $cpu =~ (.*)\((.*)x\) ]]; then
# found compressed value
value=${BASH_REMATCH[1]}
factor=${BASH_REMATCH[2]}
for j in $(seq 1 $factor); do
   conv=( ${conv[*]} $value )
done
 else
conv=( ${conv[*]} $cpu )
 fi
done

# Build .machines file
rm -f .machines

nhost=0

echo ${conv[@]};

IFS=','
for node in $SLURM_NODELIST
do
declare -i cpuspernode=${conv[$nhost]};
for ((i=0; i<${cpuspernode}; i++))
do
echo 1:$node >> .machines
done
let nhost+=1
done

echo 'granularity:1' >>.machines
echo 'extrafine:1' >>.machines

# .machines file complete

# Run your caclulation
# case name = directory name
case=${PWD##*/}

# change editor settings
EDITOR2=$EDITOR # save settings
export EDITOR=cat # prevent opening files for editing

# initialization
#cp /path/to/mstar/test_files/W2k/Si.struct ${case}.struct

# run SCF without SOC
#run_lapw -ec 0.0001 -cc 0.001 -p
#run_lapw -ec 0.1 -p -cc 0.001 -i 400 -fc 0.5 -dftd3 -min -NI
#cp -rf ../pbe ../pbe_scf

#echo "100" > kgenpt
#echo "0" >> kgenpt
#x_lapw kgen  < kgenpt


#FER=$(grep :FER $case.scf | tail -n 1 | awk '{print $10}')
#sed -i 's/0./'$FER'/g' $case.insp


x_lapw lapw1 -p
x_lapw lapw2 -fermi  -p
x_lapw optic  -p
x_lapw joint -p
x_lapw kram -p

x_lapw lapw1 -band -p
x_lapw lapw2 -band -qtl -p
x_lapw spaghetti -p

x_lapw lapw1 -p
x_lapw lapw2 -qtl -p
x_lapw tetra -p
___
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] Energy difference comparison of Win2k with GPAW calculation

[Gargee Bhattacharyya]

Dear Sir,

Thank you for your helpful suggestions. Yes, I noticed the different RMT
used for the two calculations. However, I did not set any reduction % for
both the structures, and the initial RMT for both initial and final
structures were the same. I will make sure to use the same RMT and run the
calculations.

RKmax was set to 7, but I will try with 5.5 to 6.5.

Yes, you are right, I started with GPAW calculations and took the optimized
structures from there. These were just for checking whether I can compare
the energy difference.

Okay, I will do volume optimization with Win2k and make a supercell as
you suggested and I can run with Win2k.



Thank you,
Gargee


On Wed, 11 May 2022 at 19:29, Laurence Marks 
wrote:

> Oops, I missed the RMTs -- as almost always, Peter is right.
>
> N.B., In this case the constraints mode should work. If you run into
> problems contact Peter or me -- this mode is still a bit experimental, not
> automated.
>
> ---
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> 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 Wed, May 11, 2022, 12:20 PM Peter Blaha 
> wrote:
>
>> Well, the main explanation is quite simple: Of course you have to use
>> IDENTICAL RMTs for the 2 calculations.
>>
>> I can see 1.97 and 2.23  for Ru and 1.61 and 1.17 for N.
>>
>> This is as if you would use 2 different pseudopotentials and compare the
>> energies.
>>
>> Since the N2 forces you to use small N spheres, you have to use them
>> allways. The Ru should be chosen according to the minimum distances.
>> Everything between 2.0 and 2.2 is probably fine.
>>
>> I'd even start with RKmax=5.5, but eventually increase it to 6 or 6.5
>> later to check if the total energy difference remains constant.
>>
>> In addition your cell is  "funny".  Why would you use such a rectangular
>> cell with 20 and 10 bohr ???
>>
>> And, as already mentioned by L.Marks, in WIEN2k it is much more efficient
>> to use slabs with inversion symmetry, which can be easily obtained by
>> putting the N atoms on both sides of the slab.
>>
>> You probably started with a GPAW structure, but WIEN2k has very nice
>> tools to make such calculations.
>>
>> You should first optimize the bulk Ru structure (at least the volume).
>> Using the lowest energy structure you can easily create surface slabs
>> with x supercells  (to start with I'd suggest a 3x3x2 or 3x3x3
>> supercell), and then put a N2 molecule by hand on both sides of the slab
>> keeping inversion symmetry.
>>
>> run_law -fc 10; save unrelaxed; run_lapw -min;   save relaxed
>>
>> In the same directory, you could now
>>
>> i) manually move the 2 N atoms apart and do the same steps as before
>>
>> ii) use the "constraints" option of mixer to increase the N-N distance to
>> a large (desired) value. This option even allows you to get the barrier
>> height, not just the energy of the starting and final structures. It may
>> need some fine-tuning of some input parameters, but when you watch the N-N
>> distances (and the energies) during the run you should be able to make it.
>>
>> Regards
>>
>> Peter Blaha
>> Am 11.05.2022 um 08:57 schrieb Gargee Bhattacharyya:
>>
>> Dear Developers,
>>
>> I am a new user of WIEN2k. I am using WIEN2k version 18.2 running on
>> slurm cluster. I am using the following command in my slurm job script:
>>
>> *run_lapw -NI -ec 0.0001 -cc 0.01 >& job_stats.txt*
>>
>>
>> I am trying to compare barrier height for N2 dissociation on the Ru
>> (0001) surface. Using GPAW, previously I have optimized N2 adsorbed Ru slab
>> with N-N distance 1.242 Å and 2N adsorbed Ru slab with N-N distance 2.701
>> Å using K mesh 3*6*1. I have calculated the energy difference as: *1.39
>> eV.* My question is:
>>
>> 1. When I am running scf calculation with the optimized structures and
>> comparing the energy difference, I am not getting the same. I have used the
>> same two structures and used 3*6*1 non shifted K mesh. The total energy
>> difference I found was : 0.0416 Ry = *0.57 eV*
>> I am using non-spin polarized calculations and using LDA exchange
>> functional.
>>
>> The initial structure (N2 adsorbed Ru) and final structure (2N adsorbed
>> Ru) are attached herewith.
>>
>> It would be helpful if you let me know whether I have done anything wrong
>> in the calculation so that I can compare the energy difference GPAW and
>> Win2k.
>>
>>
>> Kind regards
>> Gargee
>>
>>
>>
>> --
>> Kind regards
>> Gargee Bhattacharyya
>> Postdoctoral Fellow
>> Department of Physics and Astronomy
>> Aarhus University
>> Ny Munkegade 120
>> 8000 Aarhus C
>> Denmark
>>
>>
>>
>> ___
>> Wien mailing 
>> listw...@zeus.theochem.tuwien.ac.athttp://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>>  
>> 

Re: [Wien] Energy comparison

[Peter Blaha]

Dear WIEN2k community,
To optimize the internal coordinates by minimizing the forces it 
should not do it with SOC.


You can even switch on SO selectively for atoms, who do not have forces, 
and switch it off for atoms with free positions.



Now my question is; I want to compare a system with ferromagnetic and 
antiferromagnetic to see which one is more stable. Can I do it after I 
incorporate SOC?


Yes of course. Please use identical parameters when comparing the 
energies. In particular the Emax in case.in1 affects E-tot and should be 
chosen as high as possible and probably checked by comparing the results 
for 2 different Emax.


In addition, if you have p-semicore states of heavy elements, RLOs are 
beneficial to get better E-tot convergence.


Regards



Saludos

Pablo

___
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


--
---
Peter Blaha,  Inst. f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-158801165300
Email:peter.bl...@tuwien.ac.at   
WWW:http://www.imc.tuwien.ac.at   WIEN2k:http://www.wien2k.at

-
___
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] Energy comparison

[delamora]

Dear WIEN2k community,
To optimize the internal coordinates by minimizing the forces it should not do 
it with SOC.
Now my question is; I want to compare a system with ferromagnetic and 
antiferromagnetic to see which one is more stable. Can I do it after I 
incorporate SOC?
Saludos

Pablo
___
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