Restarting a calculation on exciting

exciting
1

Dear All

I’m running a CuS hexagonal structure to study some transport properties.
Now the exciting_smp takes a very long time execute, until the requested time expires on the CHPC.
This occurs before the convergence of the calculations could be reached.
Is there a way that one can restart a job from the end point of the previous run here on exciting?

Kindly
Thuto

2

Greetings

I’m calculating the transport properties of the hexagonal CuS phase.
Here is my input.xml file:

CuS-hexagonal-optimization 1.000000 0.000000 0.000000 -0.500000 0.866025 0.000000 0.000000 0.000000 4.330000 
  <species speciesfile="Cu.xml" rmt="1.50"  >
     <atom coord="0.3333    0.6667  0.3929"/>
     <atom coord="0.3333    0.6667  0.1071"/>
     <atom coord="0.3333    0.6667  0.7500"/>
     <atom coord="0.6667    0.3333  0.8929"/>
     <atom coord="0.6667    0.3333  0.6071"/>
     <atom coord="0.6667    0.3333  0.2500"/>
  </species>

  <species speciesfile="S.xml" rmt="1.20"  >
     <atom coord="0.0000    0.0000  0.4359"/>
     <atom coord="0.0000    0.0000  0.9359"/>
     <atom coord="0.0000    0.0000  0.5641"/>
     <atom coord="0.0000    0.0000  0.0641"/>
     <atom coord="0.3333    0.6667  0.2500"/>
     <atom coord="0.6667    0.3333  0.7500"/>
  </species>

</structure>



  <momentummatrix/>
  <boltzequ
     energyReference="efermi"
     transportDfRange="-0.5 0.5"
     transportDfSpacing="0.001"
     temperatureRange="300 300"
     temperatureSpacing="1"
     chemicalPotentialRange="-0.05 0.05"
     chemicalPotentialSpacing="0.0001"
     evOutputEnergies="true"
     siOutputUnits="true">
     <etCoeffComponents>1 1</etCoeffComponents>
     <etCoeffComponents>2 2</etCoeffComponents>
     <etCoeffComponents>3 3</etCoeffComponents>
  </boltzequ>
                                                                                                   58,16         82%

                                                                                                  1,7           Top

But I fail to obtain the desired *.OUT files.
Instead, I obtain WARNINGS.OUT with: Warning (gndstate): Reached self-consistent loops maximum.

What is it that I’m doing wrong?

Kindly
Thuto

3

Dear Thuto,

can you please submit the full input.xml?

In any case, please remember our previous discussion on the correct values for 1/3,2/3, sqrt(3)/2 etc… (here)

It seems that your run did not complete the initial ground-state calculation (it did not converge!). This can be to errors in the structure data or to the use of speciesfiles which are not appropriate.

For a more detailed analysis is always better do upload the full input file.

Your sincerely,

Pasquale Pavone (exciting team)

4

Dear Pasquale

Here is the complete input.xml:

CuS-hexagonal-optimization 1.000000 0.000000 0.000000 -0.500000 0.866025 0.000000 0.000000 0.000000 4.330000 
  <species speciesfile="Cu.xml" rmt="1.50"  >
     <atom coord="0.3333    0.6667  0.3929"/>
     <atom coord="0.3333    0.6667  0.1071"/>
     <atom coord="0.3333    0.6667  0.7500"/>
     <atom coord="0.6667    0.3333  0.8929"/>
     <atom coord="0.6667    0.3333  0.6071"/>
     <atom coord="0.6667    0.3333  0.2500"/>
  </species>

  <species speciesfile="S.xml" rmt="1.20"  >
     <atom coord="0.0000    0.0000  0.4359"/>
     <atom coord="0.0000    0.0000  0.9359"/>
     <atom coord="0.0000    0.0000  0.5641"/>
     <atom coord="0.0000    0.0000  0.0641"/>
     <atom coord="0.3333    0.6667  0.2500"/>
     <atom coord="0.6667    0.3333  0.7500"/>
  </species>



  <momentummatrix/>
  <boltzequ
     energyReference="efermi"
     transportDfRange="-0.5 0.5"
     transportDfSpacing="0.001"
     temperatureRange="300 300"
     temperatureSpacing="1"
     chemicalPotentialRange="-0.05 0.05"
     chemicalPotentialSpacing="0.0001"
     evOutputEnergies="true"
     siOutputUnits="true">
     <etCoeffComponents>1 1</etCoeffComponents>
     <etCoeffComponents>2 2</etCoeffComponents>
     <etCoeffComponents>3 3</etCoeffComponents>
  </boltzequ>

When I’m calculating the band structure, I also use the same structure, species and base vectors.
The run does complete and ground-state calculation converges well.
But once, I change the properties to calculate transport coefficient the calculation does not complete.
May some parameters in the properties sections are not compatible.

Kindly
Thuto

5

Dear Thuto,
I could not find the FULL input.xml in your last post…
This seems to be a copy of your previous post…
Could you post the full file?

Thank you.

Best,
Pasquale Pavone (exciting team)

6

input.xml (2.1 KB)
Dear Pasquale
I’ve upoaded the full input.xml in this post.
I hope it works.

Kindly
Thuto

7

Dear Pasquale

I’m sorry, here is the correct input.xml file.

Please disregard the first one.

Kindly
Thuto
input.xml (1.9 KB)

8

Dear Pasquale

That would be great.
I’ll be available on Wednesday, 29 June 2022, 6 pm.

Kindly
Thuto

9

Greetings once

I’m working on the directory CO2, diatomic molecule.
I want to analyze my results on sub-directory: CO2-test-1 using the script: $ PLOT-maxforce.py
I get the following:
Traceback (most recent call last):
File “/mnt/lustre/users/tmosuang/exciting//tools/PLOT-maxforce.py”, line 8, in
from pylab import *
ImportError: No module named pylab
How can a solve this issue?

Thanks
Thuto