FYI: superlinear speedups in GROMACS (fwd)

W Bauske wsb at paralleldata.com
Fri Mar 8 18:53:38 PST 2002


Reread what I wrote. I said 1 GB of memory....

Wes

Jim Fraser wrote:
> 
> Don't the XEON chips have a meg or two of cache?  I know they make some dual
> XEON's motherboards...maybe even quad's...have a ball!
> Also the P4's come 2X bigger L2 then the Xp's right?
> 
> IMO, any problem that fits on cache is too small and can probably be on
> calculator.  CFD is for real men... ;-)  (just kidding)
> 
> jim
> 
> -----Original Message-----
> From: beowulf-admin at beowulf.org [mailto:beowulf-admin at beowulf.org]On
> Behalf Of W Bauske
> Sent: Friday, March 08, 2002 8:14 PM
> To: rfc822 Compliance issue To: added by system POTENTIAL SPAM
> Cc: Beowulf at beowulf.org
> Subject: Re: FYI: superlinear speedups in GROMACS (fwd)
> 
> Eugene Leitl wrote:
> >
> > On Fri, 8 Mar 2002, Dominic Wu wrote:
> >
> > > Would not a single CPU with a larger L2 cache solve the problem better
> then?
> >
> > Caches are very expensive due to chip yield reasons (bad die yield goes up
> > exponentially with die size). In fact, it would make sense to drop cache
> > from dies, and go for embedded memory instead, which also removes lots of
> > other overhead, such as branch prediction and pipelining, plus allows to
> > go to very broad on-die buses and hence naturally foster VLIW and SIMD.
> >
> 
> That statement makes me curious. Do you mean embedded memory
> on chip or what? If it's on chip, how is it any better than cache?
> If not on chip, elaborate please on what you're describing.
> 
> I'd like to see a P4 with a GB or so of memory all on the chip.
> Would make an interesting node for what I do.
> 
> Wes
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list