<pre><font style="font-family: arial,helvetica,sans-serif;" size="2">Dear Paolo,<br>Here is the answer to your suggestion a few days ago, concerning the use of versions of zhpev_drv.f90<br><br>echo "4.0.4/bin-hpmpi/pw.x"<br>
/opt/hpmpi/bin/mpirun -np ${NCPUS} -hostfile nodelist /discolocal/app/Espresso/e<br>spresso-4.0.4/bin-hpmpi/pw.x -in ${inputfile} >> ${outputfile}<br> PWSCF : 34m15.70s CPU time, 36m14.28s wall time<br>
<br><br></font><font><font style="font-family: arial,helvetica,sans-serif;" size="2">echo "bin-speed/pw-1.4.x"</font></font><br><font><font style="font-family: arial,helvetica,sans-serif;" size="2">/opt/hpmpi/bin/mpirun -np ${NCPUS} -hostfile nodelist /discolocal/emenendez/Che<br>
mUtils/Espresso/espresso-4.1/bin-speed/pw-1.3.x -in ${inputfile} >> ${outputfile<br>}</font></font><br><font style="font-family: arial,helvetica,sans-serif;" size="2"> PWSCF : 35m 4.30s CPU time, 40m17.14s wall time<br>
<br>echo "bin-speed/pw-1.3.x"<br>/opt/hpmpi/bin/mpirun -np ${NCPUS} -hostfile nodelist /discolocal/emenendez/Che<br>mUtils/Espresso/espresso-4.1/bin-speed/pw-1.3.x -in ${inputfile} >> ${outputfile<br>}<br>
PWSCF : 32m15.69s CPU time, 37m13.17s wall time<br><br>echo "bin-speed/pw-1.2.x"<br>/opt/hpmpi/bin/mpirun -np ${NCPUS} -hostfile nodelist /discolocal/emenendez/Che<br>mUtils/Espresso/espresso-4.1/bin-speed/pw-1.2.x -in ${inputfile} >> ${outputfile<br>
}<br> PWSCF : 30m13.87s CPU time, 35m22.95s wall time<br><br><br>>< cdiaghg : 605.99s CPU ( 1419 calls, 0.427 s avg)<br>><br>>><i> cdiaghg : 893.34s CPU ( 1593 calls, 0.561 s avg)<br></i>><br>
>here it is: subspace diagonalization. Are you sure that the two<br>>codes were compiled in the same way? and that you are running<br>>under the same conditions? If so, please get the previous version<br>>of Modules/zhpev_drv.f90:<br>
<a href="http://qe-forge.org/cgi-bin/cvstrac/q-e/fileview?f=espresso/Modules/zhpev_drv.f90&v=1.4">>http://qe-forge.org/cgi-bin/cvstrac/q-e/fileview?f=espresso/Modules/zhpev_drv.f90&v=1.4</a><br>>then 1.3 and 1.2, and please let us know if and when you notice a<br>
>difference. If you get errors related to nonexistent routines<br>>'numroc' and 'indxl2g', just remove the line<br>> INTEGER :: numroc, INDXL2G<br>>Some changes that were needed to ensure statibility on weird<br>
>machines may have slowed down the general performances. If<br>>you have scalapack, it mighy be interesting to try it.<br></font><br><br><font size="4"><br></font></pre><br><br><br clear="all"><br>-- <br>Eduardo Menendez<br>
Departamento de Fisica<br>Facultad de Ciencias<br>Universidad de Chile<br>Phone: (56)(2)9787439<br>URL: <a href="http://fisica.ciencias.uchile.cl/~emenendez">http://fisica.ciencias.uchile.cl/~emenendez</a><br>