First AMD 45-nm CPUs May Omit Novel HK+MG Transistors

By Scott M. Fulton, III | Published February 5, 2007, 3:38 PM

In an interview with Reuters this morning, AMD's new senior vice president for technology development, Douglas Grose said that the new high-k-plus-metal-gate (HK+MG) transistor technology IBM unveiled two weeks ago might not be integrated with AMD's manufacturing process in time for its first 45-nm CPUs to be produced in 2008.

Saying the decision on when AMD will use its partner's technology depends on his assessment of AMD's roadmap, Grose explained that for now, that time appears to be "in the 45-nanometer timeframe or the 32 node application."

Last month, AMD executives gave indications to analysts, during the company's quarterly earnings call, that the 32 nm generation may be slated for the 2010 timeframe or even later.

Grose, whose hiring was announced just this morning, comes to AMD from IBM, where he was general manager of technology development and manufacturing. In 2002, when Grose was IBM's general manager for storage technology, he had been tapped as a senior executive of Hitachi Global Storage Technologies, when IBM sold a majority of its hard drive interests to Hitachi. He later returned to the fold, only to leave IBM again for the opportunity to succeed AMD's retiring Daryl Ostrander.

Ostrander's and Grose's paths have crossed before, as Ostrander was responsible for negotiating a deal in late 2004 enabling Chartered Semiconductor to produce 90 nm processors on 300 mm wafers, using technology Chartered developed in cooperation with Grose's team at IBM. The first processors produced by Chartered for AMD were unveiled last July.

Today, Semiconductor Fabtech speculated that AMD may not be as burdened as Intel with respect to integrating HK+MG transistors into its fabrication processes. AMD's existing silicon-on-insulator (SOI) wafers, the publication writes, already utilize optimum processes that Intel would have to add to its designs in order for its next-generation Core processors to integrate better with smaller processors.

Comments

View comments by with a score of at least

i think what alot of people fail to realise is what intel and AMD are about. and they are about different things. if you lump them together as equals competing for the same consumers u would be mistaken. what AMD tries to do is deliver raw processing power while intel delivers efficiency and a more compact design.

AMD caters to the people who dont mind having a massive desktop with a monstrous power supply that requires a fan so the thing doesnt explode. while intel's core due is nice and small and quiet and perfect for the business man in suit and tie to run his MS word on. AMD processors have always devilered the muscle to "mega-task" and run just about anything u can concieve on a computer.

also while people counted them down and out after intel released its core 2 duo. most people im guessing didnt hear(or care) about how AMD put a "true" quad core processor in the works. what intl did was essentially glue 2 core duos together to make their new processor. AMD is makeing a quad core built from the ground up and in the end will absolutely annihilate the core 2 duo.

Score: 0

|

"AMD is making a quad core built from the ground up and in the end will absolutely annihilate the core 2 duo"

LOL. Not quite!

While AMD's bus structure does avoid some of the bottle neck limitations and offer some advantages over Intel's bus structure, AMD is no where close to annihilating anything.

You might want to read: http://www.tomshardware....e_quad_cores/index.html

And then after all is said and done and you can actually find PC software to take advantage of the multiple cores, you might then go buy one of IBM's Power and AIX based RS6000 systems that can actually take advantage of the multiprocessing environment.

Score: 0

|

LOL.

Because AMD is not going to immediately incorporate a concept just announced several weeks ago is evidence that they are falling further behind???? Is this how its interpreted in the high schools around the country?

Without knowing the failure rates, and other contingencies, AMD would be foolish to immediately attempt to implement a conceptual advance.

The relationship between Intel and AMD will continue to be a roller coaster ride. And anyone who does not understand this or who is surprised by this does not understand the premise of the value pricing model.

If you want to complain or whine about either company, you would do well to become familiar with the economic model to which they both subscribe. Once you understand this, then yiou can better begin to understand both the realities of the market, the market forces, and also the whining and posturing of the various players.

Nothing that has happened in this market relationship thus far has been a surprise. But the ignorance of so many who expect fantastic changes in the market development has become just a bit old.

Score: 0

|

Funny how the same people that were convinced that AMD was invincible, also think HD-DVD is also invincible...

Score: 0

|

Simple:

The Underdog complex.

The front-runners equate to stagnation (logically or not), so we root for the one's how will spark competition, and hopefully, innovation.

Score: 0

|

AMD

F.a.l.l.i.n.g B.e.h.i.n.d A.g.a.i.n!

Score: 0

|

It certainly appears so. I remember their response to Intel's announcement of the Core and Core 2 lines... "We believe our technology stands on its own," or something close to that.

Well, guys, you're losing.

Score: 0

|

"In an interview with Reuters this morning..." - this is what I'm looking for.

Score: 0

|

Whoops. Thanks for noticing, that link should have been included. I've added it above, also here:

http://today.reuters.com...=BizArt-C1-ArticlePage1

-SF3

Score: 0

|

Mark Russinovich on MinWin, the new core of Windows

The next version of Windows three years hence will likely build onto a significant architectural change implemented in Windows 7 and Server 2008 R2.

Security firm: Windows patches not responsible for 'Black Screen of Death'

On second thought, maybe that access control list thingie with the lockdown something-or-rather didn't trigger an alleged, perhaps non-existent, pandemic.

My Windows 7 confession (and why you should confess, too)

I've held back the real reason for sticking with Windows 7, even as, gulp, iLife calls me to go back to the Mac.

Apple settles with Psystar except for 'circumvention devices'

The fracas with the Florida clone computer maker might have ended today had Apple not have muddled the issue over a cheap piece of Psystar software.

Comcast deal for NBCU is about content, not broadband

Although Comcast is certainly America's largest broadband provider, at least for PCs, in most regards, today's deal with GE may not impact the Internet at all.

Google begrudgingly adjusts news crawling for paid publishers

If publishers want to make readers pay for news content, and thereby drive down its popularity and Google ranking, the company says, they can just go right on ahead.

Fee or free? Murdoch, Huffington square off over the cost of Internet news

Participants in an FTC workshop yesterday witnessed the two extremes of the Web news publishing debate, still centered on the issue of long-term profitability.

Microsoft denies latest 'Black Screen of Death' claims

After an anti-malware producer announced a fix to what it says is a swarm of recent KSoD problems, evidence of the swarm itself has yet to turn up.

Latest Firefox 3.6 beta fixes 133 bugs, promises faster page load times

A once-sluggish beta testing process has kicked into overdrive, with astonishing success at finding serious bugs. Will Mozilla be able to fix all the others in time?

Confirmed: Office 2010 to ship in June

Two weeks after Microsoft had been expected to draw a clearer roadmap for its principal applications suite, it's finally ready to commit to the end of H1.

New EU antitrust commissioner will oversee Microsoft, Oracle+Sun, Intel issues

As one of Europe's most prominent politicians shifts positions in January, her replacement remains a question mark over technology's biggest issues.