Dear Axel,
I have been wondering when it makes sense and when it is adequate for getting an accurate cluster expansion to use the option -2d with mmaps or maps respectively.
Since I am working on long unit cells with a:c ratios of 1:5 I had considered it a convenient option taking notice of the fact that the relaxations of supercells along the c-axis were quite difficult to converge.
But I am not quite sure if the accuracy of the cluster expansion would suffer from the lack of the energetics of these supercells, since I am guessing that some longer range ECIs along the c-direction are not taken into account otherwise.
Could you shed some light on this issue?
Thanks in advance.
Best, Siau