Mmaps not Recognizing Converged Structures

Alloy Theoretic Automated Toolkit (ATAT)
1

Hello!

I am currently trying to build a CE with ATAT for CO on Co(0001). My lat.in and vasp.wrap files are attached. I’ve used mmaps (since I have three species in my calculations, albeit C and O are never found independently of one another), and an in-house script has been used to automate job submission to our cluster management queue.

The program has been running for a few weeks now and appears to run well; output structures look good to me and of those that have converged an ‘energy’ file is created as expected. However, In my parent directory, gs.out and gs_str.out files are empty and I have the following message in map.log:

Maps version 2.86
The internal database of structures extends up to 0 atoms/unit cell, see predstr.out
Among structures of known energy, true and predicted ground states agree
No other ground states of 0 atoms/unit cell or less exist.
Crossvalidation score: 8.50706e+37
~

I know something is very wrong here. I have a little over a dozen converged stuctures and the mmaps code doesn’t seem to be using them. Some insight could be gained by looking at the predstr.out file (also attached). The file seems to be missing the first 10 or so structures created!

So ultimately, my questions is simply: how can I resolve this situation? Can I induce the code to recognize the missing sturctures? Is there soemthing else going on here that I’m not seeing?

Any help is appreciated, and there is no amount of dumbing down that could possibly insult my intelligence. Thanks.

2

your predstr.out indicates that many structures are marked with "e"=error and "b"=busy.
error means it has noticed error messages in the ab initio output files and is ignoring the output.
If the errors are actually benign you can delete the ??/error files and maps will read those structures.
busy means that the "energy" file has not been created yet (the file energy should contain the converged total energy from the ab initio run).

3

I appreciate your response, so thank you for that. Any idea why maps doesn’t seem to be using the first 10 (i.e. directories 1 through 10) structures, though? I can look and see that predstr.out file is absolutely correct about the structures it has recognized (errors where there’s e and still running where it says b), but I do have error-free converged directories with "energy" files in them and they just don’t seem to be reflected in the predstr.out file.

Again, thank you for your help.

4

They won’t be in the predstr.out file (because they are not predicted, they are "actual").
Look in the fit.out file - they should be there.

5

And there they are. Thanks for your reply. Looks like everything is in order!