<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.2668" name=GENERATOR></HEAD>
<BODY>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>Alessandro,</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>what
is the present "official" view on using mixed MPI/OpenMP
on</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>platforms other than IBM Power? </FONT></SPAN><SPAN
class=117463310-05072005><FONT face=Arial color=#0000ff size=2>Do you think that
with arrival of </FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>dual-cores it is more practical than </FONT></SPAN><SPAN
class=117463310-05072005><FONT face=Arial color=#0000ff size=2>ever
</FONT></SPAN><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>to start using OpenMP? </FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005></SPAN><SPAN class=117463310-05072005><FONT
face=Arial color=#0000ff size=2>I'd guess many on the list would be interested
</FONT></SPAN><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>to know the developers' </FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>view
on this.</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>>
<FONT color=#000000>c) you have a system with more than 2000
states.</FONT></FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>BTW,
are there test cases publicly available (e.g. your silicon
carbide</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>example you published the report about) </FONT></SPAN><SPAN
class=117463310-05072005><FONT face=Arial color=#0000ff size=2>which would be
suitable for </FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>using
the mixed model?</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff size=2>Many
thanks in advance and all the best,</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2>Igor</FONT></SPAN></DIV>
<DIV><SPAN class=117463310-05072005><FONT face=Arial color=#0000ff
size=2></FONT></SPAN><SPAN class=117463310-05072005><FONT face=Arial
color=#0000ff size=2></FONT></SPAN> </DIV>
<P><FONT size=2>I. Kozin (i.kozin at dl.ac.uk)<BR>CCLRC Daresbury
Laboratory<BR>tel: 01925 603308<BR><A href="http://www.cse.clrc.ac.uk/disco"
target=_blank>http://www.cse.clrc.ac.uk/disco</A></FONT> </P>
<BLOCKQUOTE
style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px solid">
<DIV class=OutlookMessageHeader dir=ltr align=left><FONT face=Tahoma
size=2>-----Original Message-----<BR><B>From:</B> cpmd-list-bounces@cpmd.org
[mailto:cpmd-list-bounces@cpmd.org]<B>On Behalf Of </B>Alessandro
Curioni<BR><B>Sent:</B> 05 July 2005 10:17<BR><B>To:</B> Axel
Kohlmeyer<BR><B>Cc:</B> cpmd-list; Lin Yi;
cpmd-list-bounces@cpmd.org<BR><B>Subject:</B> Re: [CPMD-list] [Fwd: Error
Encountered During CPMD-3.9.2 Compilation]<BR><BR></FONT></DIV><BR><FONT
face=sans-serif size=2>As already stated in previous e-mails on the list - the
use of the Mixed MPI/OpenMP scheme it is useful if </FONT><BR><FONT
face=sans-serif size=2> a) you want to run on many processors (
let's say more processors than the number of realspace planes along x);
</FONT><BR><FONT face=sans-serif size=2> b) you want to run on clustered
smp servers going outside a single smp server - when the interconnect latency
and bandwidth are much worse than the the ones for the memory
</FONT><BR><FONT face=sans-serif size=2>c) you have a system with more than
2000 states.</FONT> <BR><BR><FONT face=sans-serif size=2>Your case is not in
these category - so as explained by Axel - much better to
run in MPI only mode.</FONT> <BR><BR><FONT face=sans-serif size=2>If you
have compiled the code for IBM POWER machine in the mixed version , you
should specify:</FONT> <BR><BR><FONT face=sans-serif size=2>export
XLSMPOPTS=parthds=1 ---> to specify the number of smp
threads</FONT> <BR><FONT face=sans-serif size=2>export MP_PROCS=32
---> to specify the number of MPI threads</FONT> <BR><FONT
face=sans-serif size=2>export MP_SHARED_MEMORY=YES</FONT> <BR><BR><BR><FONT
face=sans-serif size=2>In any case - I still suggest to use norm
conserving pseudos for your calculation and test with kpoints
convergency.</FONT> <BR><BR><BR><FONT face=sans-serif size=2>Best
Regards,</FONT> <BR><FONT face=sans-serif size=2>Alessandro</FONT>
<BR><BR><BR><BR><FONT face=sans-serif size=2>Alessandro CURIONI,
PhD<BR>Research Staff Member<BR>Computational Biochemistry and Material
Science group<BR>IBM Research Division - Zurich Research
Laboratory<BR>Saumerstrasse 4<BR>8003 Rueschlikon - Switzerland<BR>e-mail:
cur@zurich.ibm.com<BR>www: www.zurich.ibm.com<BR>Tel:
+41-1-7248633<BR>Fax: +41-1-7248958<BR></BLOCKQUOTE></FONT></BODY></HTML>