There you go... Anandtech posted some news saying that some sources indicate alderwood will be 5% faster than canterwood, and grantsdale should only be as fast as Canterwood.

This, of course, doesn't sound very precise, because it doesn't state anything about DDR-II 533Mhz being used, or whatever...

But if +5% comes from alderwood, and the P4EE is faster than prescott, then will P4EE at 3.4Ghz or even 3.6Ghz in an Alderwood chipset be able to counter FX-53 at 2.4Ghz and 1MB cache, possibly with 1000Mhz HT bus?... Does Intel have anything else up their sleeve? If so, will they be in such a bad position as to use it?...

Makes us curious, doesn't it?... It's hard to resist speculating... I won't even try it...

<font color=red><b>M</b></font color=red>ephistopheles

Scotty after 3.4 may go to 1MB cache 2.

I can agree that this sounds true. You will soon have the prescott cpu that will only run on certain canterwood motherboards as of now. With the Grantsdale you'll get native support with a couple extra goodies.. maybe. Next comes Tejas cpu that will not work with Grantsdale due to LGA factor, enters Alderwood with PCI-X through out, PCI-X VGA support, DDR-II memory. In the end P4 will replace celeron as the low cost chip, 2nd comes Prescott for middle range business market-speed at a resonable cost, and finally Tejas- with a preimium for the must have computer user. I848,I865,I875 [Canterwoods], I890 [Grantsdale], I900 [Alderwood]. Also dont forget the new BTX form factor is going to come out Intel product.

So that makes 5 chipsets, 5 different processors, 2 different memory types, 2 different graphic types, 2 different ATA types, 5 different Form Factors, 3 different monitor types, seems like they really want to confuse people and this is just the desktop arena. Not to bust on Intel but AMD is doing simular stuff. Barton, 64, FX-51, FX-53 and mix in all the same factors, types etc. into the mix and you have this stew effect. 2 different manufactures with 30 different ways to get one system. Each increasing the clock speed every few months on each part of the system. In the end everything still is to fast for the software its meant to run and people just fork over the money.

Hold on there for a second! Lotsa things there!

Nossir, Tejas will be LGA775, which means it will work in grantsdale. Grantsdale <i>is</i> LGA775, as a matter of fact.

Alderwood is not specifically needed for LGA775 processors. Plus, Tejas won't be out until 2005, but it should, according to current common knowledge, be compatible with grantsdale and alderwood. Another thing: <b>PCI-X</b> is not to be mistaken with <b>PCI-Express</b>. They're both different technologies; PCI-X is a server/workstation technology that has been around for several months. PCI-Express is completely new technology and is not on the market yet.

I have seen absolutely no indication of this scenario. Tejas will not compete with Prescott! Tejas will eventually - in 2005 - replace Prescott entirely, much like Prescott will replace Northwood and Northwood made Willamette obsolete. So, no, this will most likely not be the case...

Actually, the latest codes for Grantsdale and Alderwood are i915 and i925X, respectively.

These 5 chipsets are 5 chipsets over 2 years! They'll never, ever, fight each other like you seem to suggest. Plus, ATA is in a process of evolution, as is DDR, so you can't really say the two techs that are involved (old and new) will coexist for confusion! One thing is replacing another, like it has to be in every hardware evolution. One's new, the other is old. Nothing out of the ordinary.

No, what AMD is doing is a mess in so many ways. Intel has its roadmaps relatively clear now. They are transitioning a lot of things, but they <i>are keeping things pretty straightforward,</i> unlike AMD. They are changing sockets, memory and PCI bus type, but they're not making two changes in the same component over a very short time, like AMD's "three-socket-over-a-few-months" game. That is really confusing, because it implies a lot of changes <i>in the same component</i> over a <i>pathetically small amount of time</i>. Intel is not keeping things this messy and certainly not as unpredictable. Plus, AMD, while having a technical superiority right now, is getting mixed up in its marketing policies - ratings and so on will probably create a mess, if they keep this up...

Another thing... 3 different types of monitor?... I think you mean 3 different kinds of video cards, but even so?... The only transition that is going on is from AGP 8x to PEG (x16)... so there definitely cannot be <i>3</i> different types there...

I'm very sorry for ranting here... I just wanted to clear things up a bit.

<font color=red><b>M</b></font color=red>ephistopheles

Things look promising for Prescott and tejas. Prescott is already 5-10% faster than Northwood with unoptimized software, and the diff could be even greater with SSE3 and now Grantsdale and Alderwood should provide even greater boost.

-------
Kanavit's Aquamark3 rig----><A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=1817959409" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=1817959409</A>

For me Prescott is another dissapointment. Most of the "gains" will come from software. The processor will be better able to throughput large chunks of data with minimal processing done. Without the SSE2/3 software, its just a faster Willy. Guess it's just because I see a true AI needing to be better able to deal with raw data. Oh well, the guys who pirate music and movies will be happy.

well don't forget, Prescott has bigger L1, and L2 caches physically so that's gotta account for something.

-------
<b><font color=red>K</font color=red></b>anavit's Aquamark3 rig----><A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=1817959409" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=1817959409</A>

Yes and better prefetch, but again this is mostly for error prevention and information throughput. Oh well.

Erm... You're really being the pessimist here! Wait for it to flop before saying it does. A Willy? Now that is old. And bad. Seriously?...

I remember a review of the 1.6Ghz A64 we saw on anandtech once... It didn't look too promising, and that was weeks before launch. So that didn't turn out to be a reasonable impression either, did it?...

<font color=red><b>M</b></font color=red>ephistopheles

Last time I was an optomist for a chip launch, was the willy. It's worked well since then. (being a pessamist I mean) Before the P4 was released, I was so sure it would be the best thing since sex. All that hype.

Northwood turned out to be a good processor per-clock. How did your pessimism turn out then?

<font color=red><b>M</b></font color=red>ephistopheles

Perfect. I would much rather be majorly pleased, than be right.

Have you seen anything by Eden lately? Guess I'm a little worried cause North America is not a great place right now for people from the middle east. Hope he just got out of Quebec cause it's so cold there now.<P ID="edit"><FONT SIZE=-1><EM>Edited by endyen on 01/17/04 08:17 AM.</EM></FONT></P>

Yeah... That actually sounds reasonable!

But, for competition and for us consumers, let's hope you're terribly wrong this time.

<font color=red><b>M</b></font color=red>ephistopheles

I wont, dont want to linx my luck.

That a big assumtion from intel.My guess intel run I875 on intel validate module most of them run with slower timing.Alderwood might have cache on chip for pre-fetch like R659.I any case i will be really surprise if intel can produce a faster chipset with DDR-2 that actual I875.

I dont like french test