[Pw_forum] Using q point parallelization

[Gabriele Sclauzero]

zq wu wrote:
> 
> 
> On Mon, May 11, 2009 at 3:12 AM, Huiqun Zhou <hqzhou at nju.edu.cn 
> <mailto:hqzhou at nju.edu.cn>> wrote:
> 
>     Hi, list users,
> 
>     I'm trying the manual q-point parallelization and have one
>     question to confirm. In the additional information section
>     of INPUT_PH.html of 4.1 cvs, it indicates that after running
>     different subset of irrs or qs on each machine, one need to
>     "collect all the data-file.xml.#ip.#irr files in one directory
>     and run ph.x." My question is why one needs to run ph.x again
>     instead of using q2r directly to calculate FC. 

Because in this additional run ph.x will collect the data produced by all previous runs to 
build the full dynamical matrix and write it to the files (one for each q-point) with the 
prefix specified in fildyn. These files are needed by q2r to compute the interatomic force 
constants.
I think that those files should not be produced by ph.x after partial runs (e.g. with only 
a subset of irreps) in order to avoid confusion, but maybe this was fixed in a later 
version of the CVS.

> If this is
>     neccessary, what changes should be made in the input file.
> 
> To get the dyn files, which is input of q2r to calculate FC.
> You do not need change the input file except specifying  
> "recovery=.true."  

recover=.true.

GS

> 
> 
> 
>     Thanks in advance!
> 
>     Huiqun Zhou
>     @Earth Sciences, Nanjing University, China
>     _______________________________________________
>     Pw_forum mailing list
>     Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>     http://www.democritos.it/mailman/listinfo/pw_forum
> 
> 
> 
> 
> 
> Zhongqing
> 
> CACS University of Southern California
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

-- 


o ------------------------------------------------ o
| Gabriele Sclauzero, PhD Student                  |
| c/o:   SISSA & CNR-INFM Democritos,              |
|        via Beirut 2-4, 34014 Trieste (Italy)     |
| email: sclauzer at sissa.it                         |
| phone: +39 040 3787 511                          |
| skype: gurlonotturno                             |
o ------------------------------------------------ o