[Pw_forum] Problem with parallel phonon calculation using image method

[William Parker]

Hello David,

If the phonon calculation without images is correct ("aprun -N 8 -n 128 ./ph.x -npool 4 < ph.in > ph.out"), your SCF calculation ran on 128 (not 64) processors, and you need 1280 processors to do 10 images ("aprun -N 8 -n 1280 ./ph.x -nimage 10 -npool 4 < ph.in > ph.out").  Can you verify that?

--William

On Aug 13, 2012, at 8:50 PM, David Miller wrote:

> Hello,
> 
> I am trying to perform phonon calculation using image method of parallelization.  So, the following is what I do:
> 
> (i) SCF calculation for 5-atom tetragonal PbTiO3 system. First, I relaxed the system, performed a SCF calculation using 64 processors. 
> 
> 
> Next, I performed Phonon (at Gamma) calculation using the following input:
> 
> Phonons of PVDF-beta
>  &inputph
>   amass(1) = 207.21
>   amass(2) = 47.867
>   amass(3) = 15.9994
>   outdir = '.'
>   prefix = 'pto'
>   tr2_ph = 1e-18
>   fildyn = 'pto.dyn'
>   epsil = .true.
>   trans = .true.
>  /
>  0.0 0.0 0.0
> 
> and running it as aprun -N 8 -n 128 ./ph.x -npool 4 < ph.in > ph.out
> 
> The above completed successfully and I have ph.out and pto.dyn files.
> 
> Next, I tried the image method with the same file as above but with the following command:
> 
> aprun -N 8 -n 640 ./ph.x -nimage 10 -npool 4 < ph.in > ph.out
> 
> 10 images because it has 10 irreducible represetations at Gamma point. The INPUT_PH.txt in 4.3.2 version reads the following:
> 
> ***************************
>    On parallel machines the q point and the irreps calculations can be split
>    automatically. The procedure is the following:
> 
>    1) run pw.x with nproc processors and npools pools.
>    2) run ph.x with nproc*nimage processors, npools pools and nimage images.
>    3) run ph.x with the same input and recover=.true. on nproc processors
>       and npools pools and only one image.
> 
>    During the first ph.x run the phonon code split the total amount of
>    work into nimage copies. Each image runs with different q and/or
>    representations.  The second run of ph.x is the final run that
>    collects all the data calculated by the images and writes the files
>    with the dynamical matrices.
> *****************************
> 
> However, all I get in the ph.out is the following:
>     
> ****************************************
>      Program PHONON v.5.0       starts on 10Aug2012 at 22: 0:37
> 
>      This program is part of the open-source Quantum ESPRESSO suite
>      for quantum simulation of materials; please cite
>          "P. Giannozzi et al., J. Phys.:Condens. Matter 21 395502 (2009);
>           URL http://www.quantum-espresso.org",
>      in publications or presentations arising from this work. More details at
>      http://www.quantum-espresso.org/quote.php
> 
>      Parallel version (MPI), running on   640 processors
>      path-images division:  nimage    =   10
>      K-points division:     npool     =    4
>      R & G space division:  proc/pool =   16
> 
>      Ultrasoft (Vanderbilt) Pseudopotentials
> Application 16682117 exit codes: 143
> Application 16682117 resources: utime 0, stime 0
> ********************************************
> 
> Several files are generated 
> 
> directories _ph0 through _ph9 are empty and files out.1_0 through out.9_0 are all empty. 
> 
> I wonder what am I doing wrong ?
> 
> Thanks,
> 
> David
> 
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

*********************************************************
  William D. Parker                 phone: (630) 252-4834
  Computational Postdoctoral Fellow   fax: (630) 252-4798
  MSD-212, Rm. C-215
  Argonne National Laboratory
  9700 S. Cass Ave.
  Argonne, IL 60439
*********************************************************