[ase-users] speeding up thermochemistry calculations

[Claudio A. Perottoni]

Dear ASE Users,

Thanks to Cory Hargus' suggestion of reducing kpts (see message below), 
execution times are now very acceptable even for a 512 atoms unit cell 
replicated into a 5x5x5 supercell (which is probably a too large 
supercell...). Many thanks to all whom responded to the original post!

Best regards,

Claudio


Em 21-Mar-17 16:33, Hargus, Cory escreveu:
> Dear Claudio,
>
> I saw your post regarding performance of ase thermochemistry 
> calculations on the ase-users mailing list. Because it was forwarded 
> to me, I am not able to respond directly to the mailing list, but I 
> will respond as best I can here.
>
> I wrote the CrystalThermo class for ASE, so I have a bit of experience 
> with these calculations. It sounds like your performance bottleneck is 
> in the call to the ase.phonons.Phonons.dos method. When I profiled 
> this method on the example from the documentation, it seems that about 
> half of the time is spent on computing the band structure 
> (ase.phonons.Phonons.band_structure) and the other half is spent 
> elsewhere in the dos method. Both `dos` and `band_structure` iterate 
> over the points in the Monkhorst-Pack grid in a for-loop, so there is 
> certainly room for additional numpy vectorization here, though it may 
> be quite complex to implement.
>
> It's worth noting that the use of kpts=(40, 40, 40) in the 
> CrystalThermo documentation was chosen arbitrarily and might be 
> misleadingly large. In most cases, a smaller grid will suffice, and 
> this should be determined through convergence testing. In the gold 
> bulk example, the Helmholtz energy is converged to within 1e-5 eV at a 
> grid size of (4,4,4), which will be around 1000x faster to compute 
> than kpts=(40, 40, 40).
>
> I hope this helps.
>
> Best,
> Cory
>
>
>

-- 
*********************************************************************
Cláudio A. Perottoni

Universidade de Caxias do Sul
Programa de Pós-Graduação em Engenharia e Ciência dos Materiais
Rua Francisco Getúlio Vargas, 1130
95070-560 Caxias do Sul - RS - Brazil
Phone: +55 54 3218-2607

Universidade de Caxias do Sul
IMC - Instituto de Materiais Cerâmicos
Rua Irmão Moretto, 75
95765-000  Bom Princípio - RS - Brazil
Phone: +55 51 3634-1100

http://www.ucs.br/ccet/defq/caperott/
*********************************************************************


---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus