AMD XP 2800 with Thoroughbred core!!

[Coolone]

Heres a snip translated with babelfish

AMD shows Hammer and Thoroughbred Publié the 14/3/2002 by TEAM Clubic AMD benefitted from CeBIT to officially announce the next evolution of the Athlon processor known under the name of Thoroughbred code. The principal innovation of this processor is its smoothness of engraving in 0,13 micron which enables him to release less heat and thus to make it possible to reach higher frequencies. Visually celà results in a reduction of the size of the core. This new processor should arrive as of the end of the month, at the beginning of April at the latest on the market. We could photograph Thoroughbred 2800+ beside Athlon XP in 0,18 micron what enables you to distinguish the difference in size from the core.

You can find the info Here and Here

 

[ST4RCUTTER]

Supposedly it's running at 2.2Ghz. If so, that will be one amazingly fast chip. I'm pretty skeptical though. I could be wrong, but I think 2400+ at the most will be the starting point for Tbred...more likely 2200+.

 

[zemus]

2.2 Ghz athlon, i dunno if that will be seen by us anytime soon, but heh, Anything is possible I guess

 

[sparks]

Even if AMD could do it, it wouldn't happen. It would be like marketing suicide, there is LOTS of money to be made with the 2200, 2300, 2400......2700.

 

[ST4RCUTTER]

Even if AMD could do it, it wouldn't happen. It would be like marketing suicide, there is LOTS of money to be made with the 2200, 2300, 2400......2700.

Agreed. It would be a total waste to introduce a chip that's that much faster than the top Northwoods. It would just cause Intel to accelerate their frequencies and reduce the window of use/profitability on the .13um node.

 

[Coolone]

I agree, the life of the thouroughbread would be pretty short if they released it right off the bat

 

[Mavrick007]

<< Even if AMD could do it, it wouldn't happen. It would be like marketing suicide, there is LOTS of money to be made with the 2200, 2300, 2400......2700.

Agreed. It would be a total waste to introduce a chip that's that much faster than the top Northwoods. It would just cause Intel to accelerate their frequencies and reduce the window of use/profitability on the .13um node. >>


Maybe that's AMD's strategy though, but I doubt it as well. I would expect them to release the 2000+, the 2100+ and the 2200+ as the first releases. They might skip one or two of these, but I don't expect to see anything higher than a 2300+ or 2400+ as the first releases. Can't wait though cause it should offer alot more competition to the AMD vs. Intel match

 

[RaynorWolfcastle]

Umm this is kinda like AMD demo'ing a 3.0+ GHz CPU isn't it? Like others have said, I doubt
A) it will get released any time soon
B) their yields are good enough to make more than a very small number of them

-Ice

 

[Nelmster]

Hmmm...assuming the truth, this may be a decent guideline to overclockability...

 

[Rectalfier]

It's about time. I was getting worried the thoroughbred wouldn't come out for a couple months. A nice thouroughbred XP 2500-2800, will tide me over niceley untill Hammer 4400 comes out

 

[Bluga]

<< Even if AMD could do it, it wouldn't happen. It would be like marketing suicide, there is LOTS of money to be made with the 2200, 2300, 2400......2700. >>

 

[AGodspeed]

Pretty impressive stuff.

However, what's more interesting is that this could mean that the Tbreds will run much cooler than anticipated.

Recently, ZDNET Germany proclaimed that Tbred would run only 10% cooler at the same clock speed compared to the current .18-micron Palominos Athlon XP's. Considering that the Palomino runs ~ 15-20% cooler at the same clock speed than T-birds witout any process change, this seems kinda weird (although I don't know a lot about silicon and .13-micron so...).

ZDNET goes on to say that 2500+ will be the last Tbred, since it would require an already hefty 74.1 Watts of power.

So I ask now, why would AMD show off a Tbred 2800+ processor if such a processor would require an enormous 81 Watts of power and which supposedly is very difficult to get acceptable yields on since it's on such a small surface area of 80mm^2.

A Barton (SOI) processor would obviously require less power, and since AMD already has functional SOI silicon in their brand new Claw and Sledge Hammer processors, would it not be safe to assume that an Athlon can work with SOI too, since AMD has been working on Athlon silicon for much longer than Claw and Sledge silicon?

 

[Mikewarrior2]

Agodspeed,

Hard to tell whether its actually cooler than the Zdnet Germany page or not.... HOw was the demo cooled? Was it open system? Was it Liquid cooled?

Obviously, since no one has contested the validity of the Zdnet Germany site (well, contested by anyone that counts ), there's still a wait and see as far as the cooling with the T-bred chips.


Mike

P.S. In order to match AthlonXP Heat Output/Sq mm of contact surface, Thoroughbreds need to drop overall heat output by a whopping 38%. Not sure if we're gonna see anywhere near that iwth the tbred personally....

P.S.2. maybe i'm looking into it too much, but if you look at the AMD THorougbred Press Release from Yesterday, notice no real claims to major cooling drops (Unlike XP versus T-bird over-estimated claims of 20% cooler). One would figure, if there was a sizeable heat drop, AMD PR people would be all over it in the press release.....

 

[AGodspeed]

P.S.2. maybe i'm looking into it too much, but if you look at the AMD THorougbred Press Release from Yesterday, notice no real claims to major cooling drops (Unlike XP versus T-bird over-estimated claims of 20% cooler). One would figure, if there was a sizeable heat drop, AMD PR people would be all over it in the press release.....

Very good point. In fact, I think you pretty much hit it right on the nose. There's no reason why AMD wouldn't state that Thoroughbred runs 20% cooler or more in their press releases, they did it with the Palominos and there's no logical reason from what I can see why they wouldn't do it with Thoroughbred.

That whole story about Intel opening up an Athlon and proclaiming that they were eating into their .13-micron future looks more and more credible by the day...

 

[Siddhartha]

I doubt that AMD will release the Tbred at XP2800+. But I do think that the Thourghbred will reach XP2800+ and faster speeds. This is good news because it will push Intel to be competitive which is good for everyone, particularly the Intel fanboys.

 

[AGodspeed]

I just translated some of that clubic link, and they claim that Thoroughbreds will actually be selling on the market in early April. Again, they don't metion if this is the mobile or desktop variant, but I'm guessing it's the former.

If it's the latter, I'd sure like to upgrade to my 1.4GHz T-birds to 1.8GHz Athlon XP's soon.

 

[imgod2u]

Well, you're forgetting one important thing, the model and frequency that AMD releases their chips at is relatively irrelevant. The interest is in the scalability of the core. Now if the T-Bred core is capable of scaling to 2.2 GHz at initial introduction, that's pretty impressive. How will we know this? Overclocking results when they come out and people buy them. Currently the scalability of the Northwood core on air cooling seems to be 2.6-2.8 GHz. That leaves the current Palomino core to be quite lacking as far as scalability as it seems to top out at 1.7-1.8 GHz at on air cooling. If the T-bred could scale to 2.2 or even beyond, that would be pretty even the odds a little, even surpass.

 

[Mrburns2007]

A 2.2 ghz Athlon XP would be smoking....

 

[Innoka]

I don't see the Athlon as a very exciting chip anymore. The signs are that AMD do not wish to increase the cache or front-side bus on the desktop Athlon chips, or use SOI, and are placing the new chip on their traditional value footing until their 64 bit line arrives. How much speed and overclocking headroom is this new Thoroughbred line going to have? Every new line gets launched at higher and higher wattages.
Users are condemned to noisier hot computers.

 

[MadRat]

Think about all of those 1GHz Durons out there. The prospects of swapping out old chips for a brand spanking new 2200+ isn't exciting??

C'mon, its like if we could take a Coppermine 1GHz and throw it in an old LX motherboard at full speed!

 

[Mavrick007]

<< I don't see the Athlon as a very exciting chip anymore. The signs are that AMD do not wish to increase the cache or front-side bus on the desktop Athlon chips, or use SOI, and are placing the new chip on their traditional value footing until their 64 bit line arrives. How much speed and overclocking headroom is this new Thoroughbred line going to have? Every new line gets launched at higher and higher wattages.
Users are condemned to noisier hot computers. >>


Well, we've seen the Athlon XP be cooler running than the Tbird, so what's there to say that the Tbred won't be cooler as well. I expect the 0.13micron process to make overclocking easier and you won't need such fancy expensive hsf units. Maybe you will be able to use retail hsf like the Northwoods do, I doubt it though cause AMD has had heat issues and is generally a "hotter" cpu when compared to an Intel operating on the same speed and die size.



<< A Barton (SOI) processor would obviously require less power, and since AMD already has functional SOI silicon in their brand new Claw and Sledge Hammer processors, would it not be safe to assume that an Athlon can work with SOI too, since AMD has been working on Athlon silicon for much longer than Claw and Sledge silicon? >>


Isn't the Barton just an Athlon at 0.13 and using SOI process, as well there will probably be new instructions added, but it's basically the same tech? I don't expect it to work on our present mobos cause of voltage requirements, but the Tbred will supposedly work on our present rigs.

 

[acejj26]

the T-Bred will run cooler than Palomino just due to the voltage drop that will accompany the die shrink. remember, power draw varies with the square of the voltage.....so if the voltage goes from 1.75v to 1.6v, then the power drop should be the square of that ratio, which works out to be about a 17% power drop. if it runs at 1.5 V, then the associated power drop would be 27%. couple that with a heat spreader, and cooling a T-bred will be much easier than cooling a Palomino (the heat/sq mm issue would be nullified, as would the fear of crushing the core)

just my thoughts....maybe someone could clarify the voltage for the T-Bred so we can figure out a reasonable power drop?

EDIT: i just read that the power draw of the T-Breds will be 1.65v. that means a power drop of around 11% when comparing similar clock speeds. regardless, i'd love to get my paws on a 2.2GHz T-Bred! maybe it's a good thing i didn't go with the 1.6A combo i was contemplating a few weeks ago

 

[Mikewarrior2]

AMD is Demoing a 2.2ghz Thoroughbred... Nowhere in any kind of roadmap is this particular processor coming in hte next few months.... WHat does that mean? It could be exotic cooling? It could be hardcore air cooling? It could be a very handpicked sample.....


acejj26,

Thoroughbred has no heat spreader. Heat/sq mm is 1.5x that of an XP chip.



Mike

 

[Mikewarrior2]

Maverick007,

Yes, it *should* be a little bit cooler, but combined with the much smaller contact surface, will probably prvoe to be much harder to cool.



Mike

 

[vash]

This chip is probably the "best" yield they can do with current silicon. That means the chip will easily have a staggered release schedule that will have multiple processor speed lines. At the time of its debut, 2.2 may or may not be the fastest and there will defintiely be chips in between.

Can't believe that I'm already looking at upgrades for my rig(s).

vash