Alot of the "hit" with multi processors depends on how the system is
designed.  Especially the memory subsystem since you will go outside the L2
cache alot.

I'm still partial to Compaq's design, especially the more obscure things
like the ability to split IO requests across more than just one
processor...things like that.  But the memory subsystem on Proliant servers
is a lot better than some I've seen, especially the Dell PowerEdge boxes.
When running Prime95 on multiple CPU's on the same machine, the Compaq's see
far less a degradation than the PowerEdge boxes, or even IBM's servers
(which are actually pretty nice, sad to say).  I haven't had a chance to
test on any HP SMP boxes, but from my previous experience, the NetServer LX
or LS boxes at least don't really do anything special for the memory, so I'd
expect a larger "hit".

As for the off the shelf dual CPU motherboards, don't expect much.
Actually, from what's been reported, the Abit BP6 dual Celeron board doesn't
seem to do that bad, considering it's generic quality.

I'll be a bit interested to see how Athlon SMP systems fare, how well they
scale.  Another thing that will be interesting is to see how much
improvement Coppermine chips will do with the on-chip cache and with the
133MHz bus speed.  Non-cache memory access should be a lot faster
especially, with RDRAM's faster speed compared to SDRAM.  Otherwise though,
it's the same PII core that's been around for ages...don't expect *too* much
increases.

The new 820 and 840 (whenever those come out) chipsets should make generic
dual processor capability a bit better too, I hope.

> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Michael
> Oates
> Sent: Thursday, October 14, 1999 5:03 PM
> To: [EMAIL PROTECTED]
> Subject: Re: Mersenne: Mprime in dual cpu setup?
>
>
> Hi,
>
> I have a dual PII 560Mhz with NT and get similar results to you, with both
> running an exponent M8332549 ish I get times of .220 for each processor,
> with just one running it goes to .196
>
> Also, if one is doing an LL test and the other factoring the LL test will
> run at .196

_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

Reply via email to