No ground state output

Alloy Theoretic Automated Toolkit (ATAT)
1

I have a system with three elements, one of which is not varying:
lat.in:


3.59900000000 3.59900000000 3.59900000000 90.00000000000 90.00000000000 90.00000000000
1.00000000000 -0.00000000000 0.00000000000
0.00000000000 1.00000000000 0.00000000000
0.00000000000 0.00000000000 1.00000000000
-0.00000000000 -0.00000000000 -0.00000000000 Al,W
0.50000000000 0.50000000000 -0.00000000000 Co
-0.00000000000 0.50000000000 0.50000000000 Co
0.50000000000 -0.00000000000 0.50000000000 Co

After running mmaps for a while I’ve got 14 more or less decent results (in the sense mmpas reported acceptable Crossvalidation score: 0.0424162).

Nevertheless, both gs.out and gs_connect.out are empty which I interpret that mmaps could not find any ground state. This sounds a bit strange as the system is effectively binary where ground state must be easily found for any number of structures. I thought the problem is tolerance, but running mmaps with any tolerance did not help at all.

What can be done to solve (avoid) the problem?

I attached my energies and structures.

Regards,
Alex

2

This is a pseudo-binary, your can run it with maps (mmaps not needed). maps will show you that there are ground states.

Alternative: mmaps can also be made to work. In the atat/src/makefile add
-DDEGE_GS
on the CXXFLAG line
and do
make clean
make

This will turn on a part of the code that notices that a multicomponent system is actually of a lower number of species and will project it only a low dimensional space before doing the cluster expansion.
(I’ll probably turn this on by default in a later version.)

3

Thanks a lot, it helped!

4

Hello
I am working on a system that contains two elements and a vacancy. In this way, I am trying to get cluster expansion for multi-component systems with 3 components using mmaps.

I am interested in a lower concentration of vacancy, so using crange as an option with the constraint on vacancy as 1*Vac<=0.15

After running mmaps for a while I’ve got many structures with the above constraints. Among those structures, I performed ab-initio calculations for some of the structures (roughly 100 structures) and ran mmaps code.
In the map.log file, I could see the below

The internal database of structures extends at least up to 9 atoms/unit cell, see predstr.out
Among structures of known energy, true and predicted ground states agree
No other ground states of 9 atoms/unit cell or less exist.
Crossvalidation score: 0.0121353

So, I believe I am getting acceptable Crossvalidation score.

Nevertheless, both gs.out, gs_str.out, and gs_connect.out are empty and I interpret that mmaps could not find any ground state.

I am already using the modified version with -DDEGE_GS on the CXXFLAG line.

I eagerly look forward to your suggestion. Thank you in advance.

With best regards
Dhanshree