90 NM not before 2005

[juin]

http://www.eet.com/printableArticle?doc_id=OEG20030129S0049

Early 2004 at the best.Tool issue as some have say

Just next to the lab and the bunker you will find the marketing departement.

 

[Schmide]

<A HREF="http://www.eet.com/printableArticle?doc_id=OEG20030129S0049" target="_new">linkerazation</A>

Dichromatic for your viewing plesure...

 

[Willamette_sucks]

PC Magazine has a good article on CPUs and roadmaps for the future. Its in the February 2003 issue.
According to it, AMD (and presumably Intel) will be going 90nm in 2004. Both companies will go to 65nm in 2005. And intel (presumably AMD also) will be moving to 45nm in 2007, and 32 in 2009.
Width in nm of the transistor is not the only thing to take into account when talking about power usage and heat.
When both companies move to 90nm, they will also be shrinking the gate and the dielectric, which serves as an insulator to prevent the current from leaking from the gate into the pipe.
No doubt even more changes will be made to additional components (ie gate, dielectric, source, drain) when both companies move to even smaller nm processes.


I dyed my body green, branded an "n" on my chest, and stood on a box in the middle of Time Square for 3 days shouting, "ATI is the Anti-Christ!" to boost sales after nVidia had a losing quarter.

 

[eden]

This seems to be bad news for Intel and AMD.
I honestly do not know how both can hold through.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden.

 

[Schryver]

This is indeed troubling...

Does this mean I'm not going to get my Prescott for Christmas?? :-D

Seriously, though, I'm planning on building a new system when 90nm comes out... this kind of puts a damper on those plans...

So what are Intel and AMD going to be doing for all of 2004 if they're not moving to 90nm? That's what I want to know...

 

[sonoran]

Does this mean I'm not going to get my Prescott for Christmas?? :-D

Don't bet against it.

Though some chip companies such as Intel will almost certainly disagree, the panelists' comments echo a growing sentiment that chip makers will be forced to shift to a new technology node ever three years rather than every two years as many have managed to do since the 1990s.

Note that they're specifically mentioning one company that might not be content to slow down and stay in the pack. That's about all I can say...

 

[vacs]

AMD (and presumably Intel) will be going 90nm in 2004.

According to AMD latest roadmaps, they will release a 0.90 micron Athlon64 Q4 2003, not sometime in 2004. Of course, no one really believes it...

 

[juin]

Presscott is ready but trouble is meet at yield like NW on 130NM was introduce with good yield at the start.

90 NM as been delay in 1H of 2004 i expect maybe Q3 or early Q4.A64 as been delay up to H2 2003.Opteron will meet there date of release.Barton will be release in 2 week.A month or 2 for 800 FSB and HT

Just next to the lab and the bunker you will find the marketing departement.

 

[Schryver]

Hmm. Well, the way things seem to be working out... I see Intel releasing 90nm near the end of the year, in <i>very</i> limited quantities. Read: <i>very</i> expensive to get one. That or Q1 2004.

Even though AMD was supposed to release the 90nm A64 round then (says so on their roadmap), I see that being delayed at least half a year, since the 130 micron variant has been delayed by at least that much.

So assuming Intel can bring the Prescott to market in any kind of quantity, looks like they'll be the first to 90nm...

 

[juin]

Intel say 2H they have state 5% of the production in Q3 will be 0.90 and 1X in Q4.Intel go on 3 transistion on the same time strained silicon smaller node and 300 mm wafer all thing must work well

Just next to the lab and the bunker you will find the marketing departement.

 

[Schryver]

Yeah, just found a couple of articles that reiterated what you've been saying. They say intel will bring Prescott to market by the end o' the year. yay.

<A HREF="http://www.theinquirer.net/?article=7486" target="_new">Click me 1</A>

<A HREF="http://www.theinquirer.net/?article=7532" target="_new">Click me 2</A>

Seems all this talk of 90nm not going 'mainstream' till 2005 doesn't apply to Intel (and presumably AMD) 'cause they've got loads of cash to throw at the problems presented by 90nm. While that sucks for the little guys, I'm happy so long as I get my Prescott (or AMD equivalent).

On a side note, is the Prescott core still socket 478? And will it be used with the Canterwood chipset?

 

[juin]

It can be delay again if AMD dont bring any competition to intel

Just next to the lab and the bunker you will find the marketing departement.

 

[phial]

well, it seems like changing the actual chip design and makin other modifications will play more of a role in future CPU's than the increase in clock speed as we have seen in the last few years

we all know that an athlon is more effecient than a p4, clock for clock. so once we reach the *MHZ BARRIER* this factor will be really important

 

[imgod2u]

well, it seems like changing the actual chip design and makin other modifications will play more of a role in future CPU's than the increase in clock speed as we have seen in the last few years

The two are not neccessarily independent. Prescott will have over 100 million transistors as I recall. That's close to twice the transistor count of the P4. And that's a pretty big die size unless you shrink it down. Not only do you have to consider production costs, but heat generation even at moderate clockrates (moderate for the P7 core, which is around 3.2-4.0 GHz) not to mention the wattage needed.

we all know that an athlon is more effecient than a p4, clock for clock. so once we reach the *MHZ BARRIER* this factor will be really important

Again, the two aspects are not neccessarily independent. While the Athlon can achieve a higher per clock performance, the P7 core design will have a higher "MHz Barrier" as you put it than the K7 or even K8 core has. So the question is not whether one will have a higher "MHz Barrier" than the other or whether the K7/K8 core will achieve a higher average IPC (in modern software) than the P7 core implementations, but a combination of the two when they hit their respective "MHz Barriers".

"We are Microsoft, resistance is futile." - Bill Gates, 2015.