[gmx-developers] GROMACS performance on Opteron Dual Core processors!

Florian Haberl Florian.Haberl at chemie.uni-erlangen.de
Mon Sep 25 18:33:47 CEST 2006


Hi,
On Monday 25 September 2006 15:15, Christian Burisch wrote:
> Kazem Jahanbakhsh wrote:
> > Hi everybody,
>
> Hi Kazem,
>
> > Dual Core platform. From your benchmark table we saw that for
> > single Opteron processor we have a performance about 393
> > picoseconds of simulation time per day (by Intel 8 compiler) and
> > for single Opteron-DC we have  performance about 450 instead of
> > ideal expected performance about 2*393 (by gcc 3.4.3 compiler) .
>
> "N" means number of cores, not processors.
>
> > According with these results and by considering cost per
> > performance ratio it seems that Single-Core architecture is better
> > than advanced Dual-Core architecture.
>
> No. Dual core architecture is superior, and as David already pointed
> out (and as was already posted on this list) the latest Intel Core 2
> Duo Xeon 5100 "Woodcrest" seems to beat the Opterons slightly.

The market for cpus is always in fast movement, today Woodcrest (posted on the 
mailing list some weeks ago) is faster, but noone can say if the new opterons 
(better socket, perhaps quad cores) is faster, or if even Cell or Clearspeed 
will be the next big number.


>
> Regards
>
> Christian
>
>
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.

Greetings,

Florian

-- 
-------------------------------------------------------------------------------
 Florian Haberl                        
 Computer-Chemie-Centrum   
 Universitaet Erlangen/ Nuernberg
 Naegelsbachstr 25
 D-91052 Erlangen
 Telephone:  	+49(0) − 9131 − 85 26581
 Mailto: florian.haberl AT chemie.uni-erlangen.de
-------------------------------------------------------------------------------



More information about the gromacs.org_gmx-developers mailing list