<br><br><div class="gmail_quote">On Tue, Mar 3, 2009 at 9:35 PM, Paolo Giannozzi <span dir="ltr"><<a href="mailto:giannozz@democritos.it">giannozz@democritos.it</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im">Gabriele Sclauzero wrote:<br>
<br>
> I think that CG style diagonalization has not been parallelized<br>
> like the Davidson one, so these lines at the beginning<br>
> of your output sound strange to me.<br>
><br>
</div><div class="im">>> > a parallel distributed memory algorithm will be used,<br>
>> > eigenstates matrixes will be distributed block like on<br>
>> > ortho sub-group = 2* 2 procs<br>
<br>
</div>they aren't strange. This is for "subspace diagonalization",<br>
which is not directly used in CG diagonalization, but it is<br>
used in i) calculating starting wavefunctions, and ii) rotating<br>
wavefunctions before starting a new diagonalization (it is<br>
equivalent to first-order perturbation theory and improves<br>
quite a bit the subsequent diagonalization).<br>
<br>
Anyway I verified that the problem with charge is present only<br>
in parallel execution, both for CG and Davidson, irrespective<br>
whether subspace diagonalization is parallel or not<br>
</blockquote><div><br>Is there any work-around now? Parallel execution is indispensable for large systems.<br> </div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
Paolo<br>
<font color="#888888">--<br>
</font><div class="im">Paolo Giannozzi, Democritos and University of Udine, Italy<br>
_______________________________________________<br>
</div><div><div><span id="q_11fcc9e3e86e682a_5" class="h4">- Show quoted text -</span></div><div class="h5">Pw_forum mailing list<br>
<a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
<a href="http://www.democritos.it/mailman/listinfo/pw_forum" target="_blank">http://www.democritos.it/mailman/listinfo/pw_forum</a><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>------------------------------------------------------------------<br>Research Fellow,<br>Dept. of Physics, National Univ. of Singapore<br>10 Kent Ridge Crescent, Singapore, 119260<br>
Tel: +65 6516 7844<br>------------------------------------------------------------------<br><br>