The P4 is a mistake... Tualatin is the real deal.

[rcf84]

P3 core = 6 years...
1.13ghz = I'll make sure its above 2ghz
0.13micron = Overclockers Dream ;*)
256-bit cache = more data can be transfer
SSE2 = Great to add-on to a good cpu
166mhz fsb ? = I dont know much about this one
Well thats all i know.

Tualatin Should be called the P4. Not other crap intel cpu "Pentium 4"

My p3 broke :*( Well Tualatin is coming :*)

 

[rcf84]

Hey tbirdinside maybe im right here. Your afraid that the Tualatin could beat the athlon. Well this could hold up AMD till northwood comes. Well im right. Prove me wrong i dare you...

My p3 broke :*( Well Tualatin is coming :*)

 

[Guest]

No. You are very wrong. First of all Intel will only go as high as 1.26GHz cause if they went any higher 1.26PII would be challenging the higher end P4's like 1.7 and 1.8GHz P4's And 2 the P4 may one day become a good CPU. A few things need to happen. 1. SSE2, period. 2. A DDR P4 chipset with Single Channel PC3200 DDR (200Mhz. Clockspeed)

P.S. 1.26GHz it self may challeng 1.5P4!

 

[ksoth]

Isn't the Tualatin just the .13 miron PIII? Being that it is, are you sure they are going to add SSE2 to it? I mean, to me, that seems too drastic of a design change to still be the PIII... The real deal is gonna be .13 micron Northwood Pentium 4, as Intel want's that to be theri flagship processor. Pentium III will kinda turn into the Celeron of today. Basically, at .13 micron, Pentium III cores will be puny, and I mean PUNY, unless they do make design changes to the core and incorporate things like SSE2. They'll be so small that so many will fit on their wafers. And I believe Intel is gonna be increasing their wafer size to 300 mm soon as well, further reducing cost. At .13 micron, you'll see Pentium 4s at the prices of todays Pentium III, and .13 micron Pentium IIIs at the price of todays Celerons. No way is Intel gonna make the Tualatin a kick ass processor, just like what they did with the Celeron.

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[Guest]

well, hopefully the P4 will evolve into northwood, which may be halfways decent....you never know

tualatin looks like it could be great, but i can see intel pulling the ol' shifty and deliberately slowing it down.

never mind, we can always speed it up again

___________________________________________

Smoking is a major cause of statistics.

 

[Guest]

I agree the P4 is obviously the best intel has to offer right now...


So let me get this straight.
You think the new p3 will outperform its big brother the p4

i doubt intel is that foolish

I agree with Ksoth that it will just be the lower end processor and take off where the celeron will leave off

-- They have found a way to harness the power of a thunderstorm and expell it with great force!--

 

[ksoth]

But then again, I may be wrong... I mean, the Celeron is basically a Pentium III without the L2 cache, correct? There is no way that the Pentium III is gonna get it's cache taken away. So, on a clock for clock basis the Pentium III will wip the "current" Pentium 4. Will the Northwood P4, on a clock for clock basis, be much better than the current Willamette P4? If not too much better, the only way Intel can get away with selling Pentium IIIs is by intentionally leaving the clocks slow. P4 will probably hit 2.0-2.5 GHz by the end of the year, and I bet they'll keep PIIIs at 1.5 or so, just to make sure PIII will not outperform P4 (kind of like how the Celerons outperformed the more expensive PII when they first came out, which was stupid. They don't wanna do that again.) But, if Northwood P4 kick total ass and are completely superior on a clock for clock basis as the current P4 and PIII, you'd probably see both P4 and PIII hitting the same clock speeds of 2+ GHz by the end of the year. It's doubtful that Celerons will keep going, as there is not poing for Intel to have 3 different processor types.

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[Guest]

First of all the Celeron does have 128KB of L2 Cache, half of the PIII's. The only other difference is that all but the newer 800MHz Celeron has a 66MHz Bus, 800 has 100 FSB. I doubt very much that Tusaltin will have SSE2. All I've heard about it is that they are just giving it a die shrink just to make it be able to achieve higher clock speeds, and also they will NOT cut the PIII's L2 in half. If they're going to do that just get rid of the PIII and keep Celeron. And yeah there's no way in hell Tulsatin will out perform the P4. That would be suicide. I mean come on look at all they are having to do to get the P4 to succed. I truthfully cannot see how Northwood would be any better except for the fact that any DDR chipset most likely for P4 would be for Northwood and not Williamette. I mean all they're doing to it is changing it to a 0.13 micron process and MAYBE giving it L3 cache.

 

[Guest]

tualatin not having SSE2???

???????????????????????????

???

___________________________________________

Smoking is a major cause of statistics.

 

[ksoth]

It was the first Celerons that had no cache. In the slot 1 form factor, it was the same as the Pentium II/III, but just missing the external cache chips, right? Only when Intel moved to the Coppermine core did they include cache on the celeron, and wasn't it just recently with the "Celeron 2" that the cache level was 128 KB? Cuz the Duron, with its 64K L2 cahche or whatever truly spanked the Celeron, so Intel basically had to do it...

Well, apparantly the Wilammete P4 is a total hack job on the real core, and Northwood is supposedly going to have many improvments. I think the Northwood will have a larger L1 cache as well as the implimantation of an L3 cache, and some improvements to the FPU and other parts of the chip as well. That's what I've read from a couple of places, but then again that may all be just speculation. I only think L3 cache wasn't included in the Wilamette P4 because it would make the die too humungous (already it's huge), further hiking up the price and making even fewer people want to buy it. Intel would be retarded to keep the Northwood the same as the Wilamette, sans die shrink.

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[Guest]

I really don't think the PIII, even at .13 micron, is going to get much past 1.3GHz. Intel was only able to get 400Mhz out of the Coppermine (KATMAI maxed out at 600MHz). PIII is almost dead, even with a shrink. Athlon will hit at least 1.7GHz at .18 micron and probably 2.5GHz at .13u.

 

[Guest]

Yes, the first Celeron's had no cache. This is the Celeron's history:

First Celeron (Don't Know core name): 266MHz-300MHz I think, no L2 cache, Slot 1
Celeron A (Mendocino): 300MHz-533MHz. 128KB L2 Cache, running at Core Clock, Socket 370 PPGA package, don't know if there was a Slot 1 version.
Celeron B (Coppermine128): 533MHz-800MHz, 128KB L2 Cache, running at core clock, Socket 370, FC-PGA package.

Yeah I've heard the same rumors about Northwood, but the only certain thing is the die shrink. The most likely change is increase in L1 cache, and next most likely L3 cache. It would not be that suprising for Tusaltin not having SSE2. Like I have said allready Intel will make certain that Tusaltin won't compete with Williamette, and thus not go much higher than 1.26GHz. They will do all they can to make people think the PIII is dead, but it really isn't. Tusaltin could make it up there but Intel just won't do it in my opinion. I will say one thing about Athlon, AMD has done a terriffic job in making a Core that performes just as good if not better than it's competitors, and a core that will last a long time.

 

[Guest]

Your comment on Tualatin not competing directly with Willamette is interesting Here is an excerpt from a RUMOR (I EMPHASIZE RUMOR as in not confirmed but interesting as a topic for discussion as a possibility) posted on the Reg.:
"Second to lastly, but certainly not leastly, we hear that Intel is on the verge of restricting Tualatin technology, at least on its Pentium III family, to mobile processors only. This could be a blow for chipset manufacturers who think differently, but it won't be the first time such a thing has happened."

 

[rcf84]

Williamette sucks... the Tualatin is Intel's only hope left. OEM's are ready to switch AMD even Dell said there ready for the move. If the 1.13ghz Tualatin doesn't beat the Athlon 1.2ghz Intel is doomed...

My p3 broke :*( Well Tualatin is coming :*)

 

[Guest]

RC,
Although I agree with you that Wilamette "Sucks", I can't agree that Intel is doomed by any stretch. They are WAY to big for one bad product cycle to kill, and my prediction is that they will recover some of the "lost face" that this year has cost them. But, the misteps they have made this year have certainly opened the door for AMD to increase their market share. I think the days of no credible competition in the PC processor arena are done for good.

But what I am really interested in is the upcoming battle over 64 bit processing. Intel and AMD have taken two completely different paths on this one....Intel is sacrificing backward compatability/32 bit performance while AMD is going for a chip that supposedly will handle 32/64 bit performance equally well. If AMD can deliver, this could prove to be a huge advantage for them even IF their 64 bit performance is not quite as good as the Intel solution. They could provide a superior "transition" product for the move from 32 bit to 64 bit applications, allowing companies to start the move to 64 bit where they need it without having to throw out 32 bit applications that are performing just fine. This could accelerate their acceptance in the corporate arena, something that has proved to be difficult for them up to now and is just starting to happen. I DO NOT pretend to know how this will work out, but it sure is going to be interesting.

 

[ksoth]

You say PIII only got 400 MHz from the Katmai-Coppermine switchover, but that is like a 65% increase. Even if we were to base the next step on the same percentages, PIII should have no problems getting to atleast 1.65-1.7 GHz. Maybe higher if Intel uses silicon-on-insulator, but I think I remember reading something about how they are the only chip company that isn't going to adopt SOI... Not sure though. AMD road maps point to the Palomino .18 micron Athlon hitting 2 GHz by the end of the year. With Northwood P4s coming out (soon?), Athlons will need to atleast keep up in clock speed in order to compete.

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[rcf84]

Well Northwood better be one hell of a good processor. Tualatin is in the spot to be very popular with OEM's. So, no SSE2 and cache under 256kb could equal the end of intel.

Second off...

SiS has 2 Tualatin chipsets coming out. Also the i815ep supports the Tualatin. I see intel has no choise but to keep the Tualatin vs. Athlon. Cause the Current P4 sucks!!!! and Northwood is so far away.

My p3 broke :*( Well Tualatin is coming :*)

 

[Guest]

on the point about DDR chipsets for P4 , supermicro have a MB for sale in the UK....
<A HREF="http://www.scan.co.uk/allprice.htm#82" target="_new">http://www.scan.co.uk/allprice.htm#82</A>

M

Opinions are like arseholes .... everybody’s got one....

 

[slvr_phoenix]

Why in the world do people keep thinking that the Tualatin will have any less cache than the Coppermine? It'll have the same cache, if not more. That's pretty guaranteed.

As for the Tualatin having SSE2, that's likely to be bogus. If I were an Intel exec planning out my company's future, I would do the following:

Release the Tualatin as a DIE SHRINK ONLY. Let the Tualatin core go up to however many GHz it can. Don't put SSE2 into the Tualatin.

Release the Northwood with the improved FPU and more L1 cache, and with an improved branch prediction when running unoptimized code. And continue SSE2 support in the Northwood. Also, put that ability for the Northwood to act as a dual-processor system all by itself in, as well as give the Northwood multiple-processor support.

What will this simple but effective plan do?

The Tualatin, not having SSE2, nor an extra-large cache, nor support for high memory bandwidth (same flaws as in P3 and Athlon) will replace the P3 and Celeron as Intel's low- budget champ. It's speeds will probably get somewhere around 2GHz. Even at 1.5GHz though, it'd be a success as a budget chip if the prices are low enough. And without SSE2 support and with it's bandwidth limitations, it will never be able to compete against the Northwood for performance.

Meanwhile the Northwood with it's superior memory bandwidth capabilities and SSE2 support, as well as it's much improved scalability, will be able to have higher GHz ratings and well out-perform the Tualatin. On a clock-for-clock comparison, the Northwood MIGHT (and I strongly stress the unlikeliness of that might) not perform as well as a Tualatin. But no one will care because the Northwood's GHz will be so much higher that the best Northwood will far exceed the performance of the best Tualatin.

Then, if the Itanium doesn't go over well, just make a 64bit version of the Northwood.

As for the debate over the Itanium vs. the SledgeHammer, so far it's ALL theoretical because emulator performance DOESN'T count worth a darn and REAL benchmarks haven't been released for either chip ... as far as I know. (If I'm wrong, please put the links to benchmarks of NON-EMULATED results here.) So we can't judge EITHER of the 64bit chips until they're actually released to the general public.

-Despite all my <font color=red>rage</font color=red>, I'm still just a <font color=orange>rat</font color=orange> in a <font color=white>cage</font color=white>.

 

[Guest]

Like I said Slvr, we can't evaluate the 64 bit arena yet. BUT, you have to admit that it could be the next big fight between these two adn the differences between the two companies approach to this issue are interesting. It is going to make for an interesting discussion/debate as it draws closer.

 

[slvr_phoenix]

Yeah.

It'll almost definately be their next battleground because computers really could use the upgrade to 64bits. This whole 2GB memory limitation is annoying. Besides, as a programmer, I'd enjoy seeing all of those nifty new variable types on a 64bit x86 system. Plus, it would be interesting to see what the 64bit assembler code is like with the larger registers and all.

I'm definately interested in the next-gen chips from AMD and Intel.

I get the feeling thought that AMD is aiming their Hammer chips more at the budget-server / home user. They bridge that gap, but they do it by sacrificing themselves from being able to be really powerful servers.

Where as Intel is pushing it's Itanium in the direction of high-end servers / supercomputers and ignoring the fact that home users might want a 64 bit processor.

I could be totally wrong. That's just the feeling that I'm getting about each chip. Performance-wise though, I'd really love to see the two as single-processor systems and the two as dual-processor systems, running against each other, benchmarks and all.

-Despite all my <font color=red>rage</font color=red>, I'm still just a <font color=orange>rat</font color=orange> in a <font color=white>cage</font color=white>.

 

[ksoth]

This is one huge problem I see with the Northwood/Tualatin comings from Intel. Ok, right now you have Celerons and Pentium III, both of which can go in the same motherboards. You buy a Celeron system, and for the most part you can upgrade just the processor by sticking a high megahertz Pentium III and you will get a much better computer in the end. Now, Tualatin Pentium III and Northwood Pentium 4 will not be the same socket type and will not be compatible on the same motherboards/chipsets, mainly because Pentium 4 works off the quad pumped 100 MHz bus. So, the problem I see if Intel truly were to turn Pentium III into their budget processor, people would totally not be able to upgrade their systems. They'd still have to buy a new mobo and maybe even new RAM, depending on how budget it is. I think doing that is plain retarded. Will they come out with a hacked up Northwood that will turn into their budget processor??? Seems a bit sucky to me...

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[Nikko]

How did you figure that? Deshutes and Katmai, both .25, took the P6 core from the Pentium Pro, at 200MHz, to 600Mhz. Coppermine took it another 400Mhz. Now suddenly the Tualatin is going to break the trend and jump an astonishing 6-700MHz. NOT! Intel has already said they're not using SOI and even if they were the P6 will have trouble getting to 1.4GHz. Also, what AMD roadmap are you talking about? AMD's roadmap calls for 1.7 in the second half. Do you have a link?

 

[slvr_phoenix]

Are you serious?

The computer industry has changed socket/slot types how many times in the past ten years?

It seems like being unable to upgrade without replacing almost every internal component is the computer technician's motto.

The Northwood will use a totally new socket. The Tualatin will also use a totally new socket. I doubt that the two will EVER be able to run on the same motherboard.

But it seems pretty clear from Intel that Tualatin will be their low-cost processor and Northwood will be their performance processor.

So if you have a Tualatin and want to upgrade to a Northwood, you'll need an all new everything.

And if you own a P3, Celeron, or Wilamette, you may be totally screwed and need an all-new motherboard if you want to upgrade to a new Tualatin or Northwood.

But then how different is that from AMD's method? They did the whole socket to slot to socket mess too, and all the while tossing thoughts of compatability and upgradability to the wind ... until only just recently.

And from what I have heard their Hammers are supposed to use an all-new socket. And their .13micron upgrades for the Athlon and Duron are supposed to use new voltages. Same socket, but different voltages.

This means that if you have an SocketA motherboard right now, you MAY be able to support the new chips. Maybe. And again, you may not. It depends on how old the motherboard is. Just because it's SocketA doesn't mean that it will support the new chips though because it may not have the voltage settings needed.

My point? Upgradability is a real pain. When going with a significantly newer processor it should just be expected to go with a new motherboard and RAM, if not just purchase a whole new system. That, or you'll have to hunt for the specifics of what you can actually upgrade to and you'll have to be happy with less than the latest and greatest.

Such is life when PCs are involved.

It's a pain, but it's not likely to ever change.

Besides, the CPU isn't the only thing that should be upgraded to improve system performance anyway. Faster motherboards, memory, and video cards are always coming out. Sometimes you can severely improve your system performance without even touching the CPU itself. But if you're going to upgrade your CPU, you might as well make sure that it's using the best motherboard and memory possible anyway, because anything less would just slow it down.

-Despite all my <font color=red>rage</font color=red>, I'm still just a <font color=orange>rat</font color=orange> in a <font color=white>cage</font color=white>.

 

[ksoth]

If you read my post, I am talking about PERCENTAGE changes. A .25 to .18 micron shrink allowed for a 65% increase in clock speed for the Pentium III. So, if a .18 to .13 micron switch also allowed for the same PERCENTAGE increase in clock speed (65%), you may see Pentium III getting up to 1.65 GHz, a 65% increase in clock speed. It's probably not an exact science, but it makes more sense to me that using a percentage to guesstimate future clock speeds makes more sense than a set number. I mean, if you take it to the extremes, say you had a 10 GHz processor, so a die shrink would only allow a 10.4 GHz (400 MHz increase) chip? I don't think so. Do you kinda get what I'm saying now? If the Pentium III can only get to 1.4 GHz, that is a fatal move for Intel, considering Athlons are already close to hitting 1.4 GHz, soon to be much higher. A 400 MHz increase just doesn't seem like it has enough room to be successful, considering when 1.4 GHz Tualatins come out, Athlons will be quite a bit faster. Sure, if Intel is gonna be selling them as their budget chips that's not so bad, but by that time AMD might have a newer, better budget chip as well... Anyone know when Tualatin is scheduled for release?

"We put the <i>fun</i> back into fundamentalist dogma!"