Re: A very small cv but empty gs.out file, why?

Alloy Theoretic Automated Toolkit (ATAT)
1

I calculate it again. In the beginning, there are 5 lines in the gs.out file, but now ,empty.
but the CV is smaller like below:

"Maps version 3.04
The internal database of structures extends up to 21 atoms/unit cell, see predstr.out
Among structures of known energy, true and predicted ground states agree
No other ground states of 14 atoms/unit cell or less exist.
Crossvalidation score: 0.074697 "

I don’t know why? because in the beginning, maps found 5 ground state. and then , it discarded them ?

2

I recently corrected a bug in the convex hull code (in mmaps). Please try version 3.20 (beta version of the latest update).
Sorry for the inconvenience!

3

Hi Guys,

it seems I have stumbled into this problem, too. Let me post my output:

Maps version 3.20
The internal database of structures extends at least up to 14 atoms/unit cell, see predstr.out
Among structures of known energy, true and predicted ground states agree
No other ground states of 14 atoms/unit cell or less exist.
Crossvalidation score: 0.0630114

All gs files are empty for me, too.
predstr.out has 7131 entries and no g, no e but only u.
My chempots.out only states:

chemical potentials stabilizing multiphase equilibria
ground state indices , structure labels : chemical potentials
chemical potentials stabilizing each ground state
ground state index , structure label : chemical potentials

so it seems there is nothing found. I did run 30 structures and I checked that all individual directories contain the file energy with the correct DFT energy, they do. I ran a simple test system containing Cr,Fe in bcc from lat.in.

I do also see that at some point there are gs-files and chempots.out with values but when I include more data the files turn up empty again.

Any clues as to what is going wrong here?

Thanks very much, Jan

4

I got the same problem when I use mmaps, too. It’s my maps.log:

Maps version 3.36
The internal database of structures extends at least 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: 0.103785

All gs files and predstr.out are empty for me, too.

I have been run 100 structures, what should I do?

5

The problem is solved by adding -z=1e-5.

The detail is in https://brown.edu/Departments/Engineering/Labs/avdw//forum/viewtopic.php?f=3&t=319

6

Set ‘-z=1e-5’ didn’t work for mmaps, is there any other way to fix the problem?

7

this Bug is Reappear,In version 5.4.4
Thank you for fixing it!