[ase-users] Quantum Espresso interfaces

[Ask Hjorth Larsen]

Hi,

Am Mo., 26. Nov. 2018 um 23:30 Uhr schrieb Noam Bernstein via
ase-users <ase-users at listserv.fysik.dtu.dk>:
>
> Does anyone have any experience or suggestions for ASE interfaces to Quantum Espresso?  I'm interested in trying them out, but want to know if there are any opinions or comments about things like the built-in ASE calculator vs. other options (e.g. https://github.com/lmmentel/ase-espresso/).  I'm particular interested in things like minimizations driven by ASE's optimization algorithms (so from QE's point of view repeated single point evaluations that could efficiently start from the previous evaluation, but still able to control whether or not the previous geometry's wavefunc/density are used, because sometimes I do multiple different configs), and high-throughput like evaluation of 100's of unrelaxed configurations with a minimum of hand-tuning of SCF convergence parameters.

I'd of course use the built-in one unless there are specific reasons
to use something else. but then I'm totally biased of course.

The built-in QE can be used with the new SocketIOCalculator in
ase-3.17.0, which allows keeping the runtime information when moving
atoms and changing the cell.  The socket feature as implemented within
QE is a bit wonky as far as I could see when testing it (compared to
Siesta and Aims which also support this).  If you see any issue with
it, it will be a high priority to fix it.

Best regards
Ask



>
>                                                                         thanks,
>                                                                         Noam
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users