What is the 'Tualatin' Intel chip all about?

[Wingznut]

I guess I was referring to this statement by Mungla... ""Have you seen the price for a Tualatin? That danged thing is more expensive than a P4!"

Btw, mobile cpu's are always priced considerably higher than their desktop counterparts. Yes, even Athlons and Durons.

 

[ElFenix]

tualatin is mostly for notebooks anyway... it would be nice if you could drop one into your current s370 board but pretty much none of them support that voltage. in notebooks speed isn't the primary concern, mostly. heat and battery life on the other hand, are crucial. i did a bit of playing with a mobile duron at best buy, looked for hotspots, didn't find any, but then it was only running windows desktop and probably was powernow'ed and HLT'd. i'd like to see one decoding a dvd or a divx movie on an airline tray table.

that said, i think the tualatin is exactly what the notebook market needs: a very low power yet high-performing chip that won't burn your lap when you're actually using processing power. 1.2GHz athlon vs 1.2 GHz p3 in a notebook, i think most buyers would probably choose the intel if the price were the same, because speed is pretty equal but battery life should be much better.

 

[ST4RCUTTER]

ElFenix,

I would tend to agree. Of course at this point any AMD product in the mobile market will help their bottom line. It will probably boost their ASP too!

 

[BFG10K]

Wingznut:

So, when you guys speak of the Tualatin being faster than a P4, but slower than an AMD...

Using a bit of logic and using Tom's benchmarks. The P4 has a lower IPC than any other current processor, hence it will never be able to compete on a clock for clock level without SSE2 enhancements.

From what I can tell (clock for clock): TBird > Tualatin 512 > Tualatin 256 > Coppermine > P4.

And also the prices that it will be...

I disgree with the other guys about the pricing. I expect the processor will be significantly cheaper than a Coppermine. If nothing else, the smaller die size will increase yields.

Where are you getting this information at?

Have you seen Tom's latest Tualatin preview?

 

[splice]

<< Hmmm...I suppose Intel could make the Tualatin cheaper than the current P3...how much do you want to bet they don't?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
P.S.
Man, just looking at Newegg makes me sick...I can pick up a 1.4 Athlon for what I got my 900 for 4 months ago! >>



It's almost impossible for Intel to market chips at AMD's prices.... they have just too much overhead.

 

[ElFenix]

i wish they would use some of that overhead to do something about that building skeleton they've left in downtown austin... i drive by there twice a week or more...

 

[ST4RCUTTER]

I don't see how increased die yields will automatically mean a decrease in price. Yes, they will get more die per wafer, but they need to recoup the cost of the new process. If they (Intel or AMD) spend millions on new fabbing equipment and then lower the price of the processor they are shooting themselves in the foot. They would have to produce more than twice as many die per wafer to lower prices and not take an ASP hit that would hurt their bottom line. It would be bad business. Simply put, if you take one step forward with a die shrink and one step back with lowered pricing, you're stuck at the same margin levels.

 

[splice]

Yeah, I've heard about that (edit: partial building)... sucks.

 

[ElFenix]

i wonder if 12&quot; wafers will be employed in the tualatin? more than twice as much area per wafer, plus the shrinkage, and you're looking at 4x as many dies per wafer.

 

[Wingznut]

Pretty much the only reason that AMD prices their cpu's as low as they do, is because they have to. If they were to price them equally to Intel, nobody outside of this tiny demographic (hardcore enthusiasts) would buy them.

If the roles were reversed, AMD would be charging as much as they could for their product... Just like any other good corporation.

As for that Tom's article... Considering that he has no idea what that cpu is, it's hard to take that article at face value. And since it appears to be an engineering sample, it's extremely possible that it's not indicative of what the final Tualatin silicon will be. Engineering samples are made at all stages of cpu development. It could be several steppings old.

As for the Overclockers link to the article written in Chinese... No telling what they've got going over there.

I've said it a bunch of times before... Just because someone posts some pretty graphs on the 'net, doesn't make it fact.

 

[CraZCaM]

<<Scouzer is right, Intel relies on the software designers to do the dirty work of optimizing their program code. Without optimized code, the P4 is left in the dirt.>>

I 2nd that Mungla

CraZCaM

 

[nam ng]

Mungla... Would you contribute in the following thread? Please, Thanks...RIGHT HERE

 

[BFG10K]

Wingznut:

As for that Tom's article... Considering that he has no idea what that cpu is, it's hard to take that article at face value.

What was wrong with it? I thought he showed the Tualatin in a very good light. If the final stepping makes it even better, that's just icing on the cake. Also it's a very good sign to have such an early version overclock that well.

As for the performance, since there are no major core overhauls, I'd expect it to perform in a very similar fashion to a regular P3. I saw nothing suspicious or odd-looking in those results.

 

[Degenerate]

i thought there were some minor changes in the tutalin? new instructions?...

What would the northwood be like? what changes? i simply cannot wait.

 

[Degenerate]

regarding the notebook area. has anyone got a A4 notebook? how does it feel? temperature?

 

[TuffGuy]

Tualitin vs Coppermine vs Tbird @ 1192mhz w/ 159mhz fsb

and keep in mind that the tualitin uses less than half as much power as a tbird.

 

[Boogak]

<< i wish they would use some of that overhead to do something about that building skeleton they've left in downtown austin... i drive by there twice a week or more... >>


Imagine how Fort Worth feels after wooing Intel to construct a new semiconductor plant here a few years ago and having Intel dump their plans all of a sudden. I think there's just a partially completed building amidst a huge construction site now.

 

[christoph83]

Tuffguy that shows the 512k cache version. It gets a little boost from that. But overall it looks like its going to be a very good chip.

 

[TuffGuy]

yeah i know, but you still have the coppermine which the tbird is supposed to blow away by &quot;40-50%&quot; clock per clock...

and it still uses over less than half the power.

 

[ElFenix]

IMHO the p6 is the best core intel ever made, and in fact is the most successful (sales). and this last revision looks to be the best yet. too bad politics will help kill it before it needs to be, ala the 486.