<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=gb2312">
<META content="MSHTML 6.00.2900.2523" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT size=2>
<DIV><FONT size=2>Hi,everyone.</FONT></DIV>
<DIV> I'm sorry to post this mail again, but it really
disturbed me and I haven't found a solution in the previous post.</DIV>
<DIV><FONT size=2> I am a newbie to gromacs and have
used NAMD for some time. I think maybe Gromacs can do something NAMD can't, So I
came into your area.</FONT></DIV>
<DIV><FONT size=2> I have compiled single precision versions
of fftw and gromacs with icc. The Cflags used were "<SPAN lang=EN-US
style="FONT-SIZE: 12pt; COLOR: red; FONT-FAMILY: ËÎÌå; mso-bidi-font-family: ËÎÌå; mso-ansi-language: EN-US; mso-fareast-language: ZH-CN; mso-bidi-language: AR-SA">-O3
¨Cip ¨Cunroll -xN</SPAN>". Now the first problem bothered me was the parallel
performance of Gromacs. It was unreasonable on a Xeon
2.4 cluster made by myself. The OS was Rocks 3.2.0 based on RedHat 3.0EL,
network was 1000Mbit ethernet. Hyperthreading was on and it was about 10%-20%
faster than off while running namd.</FONT></DIV>
<DIV><FONT size=2> The option used like
this:</FONT></DIV>
<DIV><FONT size=2> grompp -f grompp.mdp -c conf.gro -p
topol.top -po mdout.mdp -np $NSLOTS -shuffle -sort<BR> mpirun
-np $NSLOTS -machinefile $nodefile mdrun -s topol.tpr -o traj.trr -c confout.gro
-e ener.edr -g md.log</FONT></DIV>
<DIV><FONT size=2>benchmarks:</FONT></DIV>
<DIV><FONT size=2> Villin:</FONT><FONT
size=2><BR> CPUs
(Mnbf/s) (GFlops) (ps/NODE hour) (NODE
hour/ns)<BR>
1: 45.429
1.424 345.125 2.897</FONT></DIV>
<DIV><FONT size=2>
4: 46.467
1.458 352.941 2.833</FONT></DIV>
<DIV><FONT size=2>
8: 29.621 929.329
225.000 4.444</FONT></DIV>
<DIV><FONT size=2> DPPC:</FONT><FONT
size=2><BR> CPUs
(Mnbf/s) (GFlops) (ps/NODE hour) (NODE hour/ns)</FONT></DIV>
<DIV><FONT
size=2> 1:
18.722 675.081
5.781
172.968<BR>
8: 33.108
1.195 10.221 97.833</FONT></DIV>
<DIV><FONT
size=2> 16:
42.312 1.528
13.072 76.500<BR> Considering
Hyperthreading was on, while CPUs was 4, actually it was 2.</FONT></DIV>
<DIV><FONT size=2> The performance was far away from the
benchmark on website of Gromacs, and almost as same as a previous post
describing: <A
href="http://www.gromacs.org/pipermail/gmx-users/2004-June/011028.html">http://www.gromacs.org/pipermail/gmx-users/2004-June/011028.html</A>.
but I haven't seen his solution.</FONT></DIV>
<DIV><FONT size=2> What's the problem? How cloud I do? Maybe I
should close hyperthreading or recompile Gromacs or other thing?</FONT></DIV>
<DIV><FONT size=2> Please help me. </FONT></DIV>
<DIV><FONT size=2> Thanks a lot.</FONT></DIV>
<DIV><FONT size=2>Wenyu Zhong</DIV></FONT></FONT></DIV></BODY></HTML>