QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: Juan on August 18, 2015, 15:21
-
Hello!
I optimized my BiFeO3 rhombohedral structure by relaxing the crystal until 0.001 meV/A force was calculated. I ran the program on a cluster. It never showed any error for the calculation. Once I imported the .nc file to the program to see the bandstructure, the error on the attachment appears.
I looked in the FAQ and also at the forum but did not find anything about this error. A detail is that, once I hit the "Ok" option, the program opens a bandstructure of another crystal I already calculated the bandstructure. Any help about what I did wrong or what "plugin" I'm missing will be much appreciated!
Thank you!
-
This other crystal for which you already calculated the bandstructure, is it on the LabFloor when you try to visualize the bandstructure of the new crystal? If so, did you try to "uncheck" the old calculation such that only the new calculation is on LabFloor?
-
Yes, it is on the LabFloor. Yes, I unchecked the old calculation and still appears the same error.
-
Did you try to recalculate the bandstructure from the optimized BiFeO3 configuration?
Also, could you attach the script you used for optimization + bandstructure calculation?
-
No, I did not do the recalculation after the optimization. I'll do that right now.
-
I just ran your script and did not experience any problems. See attached bandstructure plot.
-
Thank you very much. Anyhow, I still need to check on that because I need to run more calculations. If it ran without problem, probably I'm missing something either a plugin or something is going on when I translate the file from the server to the computer. I really appreciate the help!
-
Sure!
A disclaimer: I only uploaded the bandstructure plot as documentation that the script worked for me. I intentionally decreased the k-point sampling and force convergence criteria in order to get the full calculation running fast. So please do not use the uploaded bandstructure :)
-
Document optimized my BiFeO3 rhombohedral construct by relaxing the crystal until 0. 001 meV/A trigger was calculated. I ran the diet plan on a cluster. It never showed any error for those calculation. Once I brought in the. nc file to program to see all the bandstructure, the error relating to the attachment appears.
I looked on the FAQ and also inside the forum but did not find anything regarding it error. A detail is normally that, once I reached the "Ok" option, the diet plan opens a bandstructure about another crystal I actually calculated the bandstructure. Any help with what I did wrong or possibly what "plugin" I'm missing would be much appreciated! http://biannoky.tumblr.com/
-
Document optimized my BiFeO3 rhombohedral construct by relaxing the crystal until 0. 001 meV/A trigger was calculated. I ran the diet plan on a cluster. It never showed any error for those calculation. Once I brought in the. nc file to program to see all the bandstructure, the error relating to the attachment appears.
I looked on the FAQ and also inside the forum but did not find anything regarding it error. A detail is normally that, once I reached the "Ok" option, the diet plan opens a bandstructure about another crystal I actually calculated the bandstructure. Any help with what I did wrong or possibly what "plugin" I'm missing would be much appreciated! To you know know more click this link http://biannoky.tumblr.com/
-
I followed your recommendation about first doing the optimization and then calculating the bandstructure for the optimized rhombohedral structure of BFO. The program finishes but I still get the same error. I really don't know what could the problem be. I've been thinking if it has to do with the displacement of the atoms after the optimization?
Thanks
-
I will need to run the full relaxation before commenting on that. Stay tuned...
-
BTW: which ATK version are you guys using?
-
That's a good question, because although we can see VNL is 2014.2 (btw - you should upgrade to 2014.3!) I suspect the calculation of the band structure is done with 13.8?
-
To upgrade, I download again the program from the page, right? I think there should be an "upgrade" option at the top bar in the LabFloor.
-
Yes, to update you download the release you want to run and install it.
Each new release usually comes with so many new features and improvements that a simple "upgrade" option is not feasible for us.
-
Thank you, Wellendorff, for the help! I'll download the update and see if that works.
-
The upgrade itself will NOT help. However, if you run the entire calculation in 2014, I see no reason it should not work. Just make sure not to save the results in an NC file that was earlier made with 13.8. If you still have problems, submit the Python script and NC file to us for further troubleshooting.
-
Hello!
I ran the calculations once more. I did it for the BiFeO3 Rhombohedral structure (no optimization) and with optimization (Optimized1 and Optimized2) of 0.001 eV/A and 0.05 eV/A respectively. Again, the same problem appears. I'm trying to upload the .nc files but they are large and the post won't let me. The one I'm interested in is the Optimized1. Was the first one to show me the error message. If there is other way to let you have the .nc files, let me know. I'll wait for your response. Thank you!
-
Hi Juan.
I downloaded and ran your script "BiFeO3RhombohedralOptimized1.py" without making any changes to it. Calculation runs fine, the bandstructure appears on the LabFloor, and I can visualize the bandstructure with no problems at all.
I suggest you try to run the calculation locally on your own machine (took 1h 40 min on my machine) to check if there is some kind of incompatibility between the ATK versions on your machine and the cluster.
-
You really need to check which version you are running, and ensure you are not saving the results into an existing file made with an older version of ATK.
-
Hey!
I just wanted to say thanks for the recommendations. I actually ran the optimizations on my computer and there was no problem. I can see the bandstructure. My computer has 14.2 version, same as the cluster. I don't think it was the version difference. Although, as Blom says, probably I was overriding some file without me knowing about. Again, thank you very much!
Best Regards
-
All in good time :) We are so busy developing new scientific features, it's hard to also do all such "simple" things :) We hope people appreciate more that we add some new calculation function, and can live with downloading and installing "manually".