Archived from groups: comp.sys.ibm.pc.hardware.chips (
More info?)
On 10 Apr 2004 02:55:21 -0700, news.yaya.bbbl67@spamgourmet.com (Black
Jack) wrote:
>harryguy082589@aol.com (Dan Irwin) wrote in message news:<2a779348.0404081744.95f199@posting.google.com>...
>> I'm thinking of buying the a high end 478 MoBo with a Prescott CPU. I
>> have been debating weather or not i will get a good balance of
>> longevity and power out of it, because it is not 64 bit like the
>> athlon64s are. A bit after thinking about this i heard that later
>> Intel will add 64 bit to Prescott. Now the question i am posed with is
>> will these chips work with the current boards? Because i don't know
>> the answer to this question i am now posing it to you.
>
>Well, as others have pointed out this is the end of the line for
>Socket 478 Pentium 4, the next jump is to Socket 775. The 64-bit
>features won't be available _at_least_ until then. Note that this
>doesn't mean that as soon as Socket 775 arrives, then that 64-bit
>Pentium 4 will too. The first generation of S775 P4 will still be
>32-bit only.
No reason that I see that the current socket couldn't be used for the
extended addresses - the pins are there on the CPU package... up to 36-bit
at least and in theory could work with a E7505 chipset. This is really a
chipset issue with the indirect effect of mbrd traces.
>I'm not sure why Intel is waiting so long to introduce the 64-bit
>features.
As already hinted I think there's a chipset mess brewing. If
Grantsdale/Alderwood have the extended addressing included or not, either
way, it's going to make it umm, awkward to segment the market, as Intel is
fond of doing. What *would* they do with Celeron then?
This is part of the elegance, both technical and marketing-wise, of the AMD
approach - all the market layer tweaking is on a single CPU package and
there are no real issues with mbrd compatibility.
> I think Intel may feel that it's not completely compatible
>with AMD yet. It's widely believed that Intel took some first draft
>AMD documents and reverse-engineered based on those preliminary
>standards, but AMD made some changes to the standards in the final
>drafts which may have caught Intel off-guard. Intel is in an
>unfamiliar location now. Until now, it drove all of the standards on
>this architecture, and therefore it was upto others to follow its
>standards. I don't think Intel is too familiar with following someone
>else's lead.
Uh-huh this is all very unfamiliar territory for "i" at least in the spin
dept.. What next?... HyperWhat?
Rgds, George Macdonald
"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??