[Pw_forum] problem while running parallel version of espresso-4.1.1

Bipul Rakshit bipulrr at gmail.com
Tue Dec 8 12:56:22 CET 2009


Dear PWSCF users,
I am using a Linux clusters, with ifort and mkl libraries.
I am running a calculation of LaMnO3 with 20 atoms. Following the the part
of the output which i am getting on running scf calculations. The strange
thing is, although it shows approximately 26 secs for each iterations, but
almost after 4 hrs it able to complete only 4 scf only.

*     total cpu time spent up to now is      8.15 secs

     per-process dynamical memory:    20.0 Mb

     Self-consistent Calculation

     iteration #  1     ecut=    25.00 Ry     beta=0.50
     Davidson diagonalization with overlap
     ethr =  1.00E-02,  avg # of iterations =  2.0

     total cpu time spent up to now is     26.64 secs

     total energy              =   -1612.92488505 Ry
     Harris-Foulkes estimate   =   -1599.80692247 Ry
     estimated scf accuracy    <      15.29925563 Ry

     total magnetization       =     0.00 Bohr mag/cell
     absolute magnetization    =    19.65 Bohr mag/cell

     iteration #  2     ecut=    25.00 Ry     beta=0.50
*
I compiled the espresso in parallel without any error. I am running this
calculations on 16 processors, and in 4 hrs it completes only 4 scf cycles.
with the mail i am sending you the make.sys file and the input file.
Can anybody tell me why the calculations taking so much time? Is there any
other flags i have to use in make.sys file to speed up the
calculations*? *Kindly
tell me if you need any other information about the machines
*
*Bipul Rakshit
Research Fellow
S N Bose Centre for Basic Sciences,
Salt Lake,
Kolkata 700 098
India
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20091208/847c7f2f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: LaMnO3.zip
Type: application/zip
Size: 6624 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20091208/847c7f2f/attachment.zip>


More information about the users mailing list