[Pw_forum] 回覆: a phonon calculation problem

geoffrey.tse geoffrey.tse at pku.edu.cn
Fri Apr 24 16:16:15 CEST 2015


First, add 


已從我的三星裝置發送


-------- 原有訊息 --------
由: 杨顶峰 <yangxunscience at 163.com> 
日期: 2015-04-24  21:20  (GMT+08:00) 
收件人: pw_forum at pwscf.org 
主題: [Pw_forum] a phonon calculation problem 

Dear,
First, add the following command to the scf input file, not the phonon input.
wf_collect = .true.,
Second, use the same number of processors for both to simulation run.

Geoffrey
Peking university, china




                 I am trying the qe5.1 to calculate the phonon dos and phonon dispersion curve of a crystal compound. First, geometry structrue and a single Energy calculation were ran sucessfully. Next, WHen i began the dynamtrxi calculation, there is a problem. 
  pesudo potential files: 
Cu.pbe-mt_fhi.UPF, S.pbe-mt_fhi.UPF, V.pbe-mt_fhi.UPF
 my cu3vs4.ph.in file contents:
phonon for cu3vs4
&inputph
tr2_ph=1.0d-14,
prefix='cu3vs4',
amass(1)=63.54,
amass(2)=50.9415,
amass(3)=32.06,
outdir='./',
fildyn='cu3vs4.dyn',
trans=.true.,
ldisp=.true.
nq1=5, nq2=5, nq3=5,
/

and it gives wrong results:
  task #         6
     from phq_readin : error #         1
     pw.x run with a different number of processors. Use wf_collect=.true. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

  so i add the "wf_collect=.true."   in the cu3vs4.ph.in files,and run the order "mpirun -np 8 ph.x < cu3vs4.ph.in > cu3vs4.ph.out", it still can not work and shows:

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%      task #         4
     from phq_readin : error #        19
     reading inputph namelist
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

how can i conquer the problem ? Please help me , thanks a lot! 
                  


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20150424/016a8df4/attachment.html>


More information about the users mailing list