> On  7 May 98 at 19:03, Michael A. Wosnick wrote:
>
> > According to my source, there is a little known limitation with current
> > Pentium II processors that allow them to recognize or address a limit of
> > 64Megs of RAM.

Sounds like the same refrain we've heard time and again about all
Pentium CPUs being limited to 32 MB or 64 MB, etc. Its the chipset
and MB (cache/tag ram), not the processor!

<snip>
>   This is a limitation in the Intel 430VX and 430TX *Pentium* (NOT
> PII!) chipsets.
<snip>
>   The limitation doesn't prevent recognition or addressing, just
> caching.  Many benchmarks show this as a drop in system performance
> of about 5%, but in a few cases the impact can be as much as 15%.
>
> > According to this source of mine, there is a new flavour of Pentium II
> > coming out soon that will eliminate this limitation, and he predicted that
> > the prices of current Pentium IIs will plummet as a result.
>
>   The "new flavour of Pentium II" is aimed at low-cost applications,
> and so is more likely to "prop up" prices for existing PII versions
> than to drive their prices down.

This "new" Pentium II reduces costs by doing away with the L2 cache
altogether! I guess it would eliminate the potential problems of
"too much" memory to cache -- no caching at all! But at what
performance cost? The May issue of PC World was not too keen on this
new Pentium II, the Celeron. In addition, it appears to be a "dead
end." Reports are that you will not be able to upgrade from the
Celeron to a "full" Pentium II without changing motherboards. Intel
is trying very hard to move everyone to Slot 1 and 2 to eliminate AMD
and Cyrix and any other pretenders to the Socket 7 throne they
vacated. Reminds me of the step back Intel made with the crippled
386SX chip.


<snip>
>
> David G
>

Just my $0.02

Jim
Jim Maki
[log in to unmask]