[Pw_forum] ph.x: Avoiding the recalculation of the band structure in distributed phonon dispersion jobs

[Karttunen Antti]

Dear Lorenzo and Paolo,

Thank you for the suggestions concerning the band structure calculations within distributed phonon dispersion jobs. I'll think about modifying the source code like Paolo suggested. In any case, I know we're not the only ones who are taking advantage of the grid parallelization since it's very useful for large structures.

Best wishes,
Antti

-- 
Dr. Antti Karttunen
Department of Chemistry
University of Jyväskylä, Finland
Tel: +358-50-3473475
WWW: http://www.iki.fi/ankarttu 

From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On Behalf Of Lorenzo Paulatto
Sent: Tuesday, February 12, 2013 8:01 PM
To: PWSCF Forum
Subject: Re: [Pw_forum] ph.x: Avoiding the recalculation of the band structure in distributed phonon dispersion jobs

The old phonon code (until 4.1 from a very superficial search I just did) used to, or at least could, work like that: you had to perform a nscf pw.x calculation with a special option and than then ph.x calculation, and keep repeating the two steps for every 1-point. 

I'm not sure if there is any overlap between the ph.x version that do not do the nscf calculation themselves and those that contain the irrep parallelism, but it may be possible to recompose the 2.

bests