Original URL: http://www.theregister.co.uk/2007/01/28/intel_ibm_highk/

Intel teaches IBM how to reveal chip breakthroughs

Long learning process

By Ashlee Vance

Posted in Servers, 28th January 2007 18:40 GMT

Were you to invent the most fundamental change to processor guts in over forty years, you'd want to tell a lot of people about it. So why on Earth did Intel and IBM reveal such a breakthrough on Friday night, as Silicon Valley went to sleep? That question sparked our interest when Intel and IBM's news statements about new transistor designs appeared together at 9 pm.

Chip makers have used silicon as the element of choice for such transistors for more than forty years. Now, because of atomic-level constraints, they've had to a pick a new element - hafnium - and a pair of undisclosed metals to get the same job done. Finding the right materials and manufacturing techniques to produce these new transistors has taken years and years of blood, sweat and tears research. How funny then that after a decade of grunting in the labs Intel and IBM simultaneously nailed the scientific breakthrough, issued press releases about their achievements at the same time and shared in their ambivalence as to whether or not anyone read about the game-changing events.

Cough

In reality, Intel's engineers did want their deserved glory, and they wanted it all to themselves. But IBM had to spoil the show.

Last week, Intel revealed its transistor breakthrough in front of a group of reporters at the company's Santa Clara headquarters, although the press were sworn to temporary secrecy about the meeting.

The company showed off the so-called High-K + metal gate transistors that have already started to replace standard silicon-based transistors in chips made via Intel's new 45nm manufacturing process. Such chips will ship later this year. [In a previous report, we described some of the fine-grained details behind the new transistors.]

Armed with the fresh transistors, Intel has promised better performing products for years to come - products that match up to Moore's Law.

In addition, Intel has claimed more than a one-year edge over rivals by adding the new transistors to its mainstream production line so quickly.

"It shows you the lead Intel has today and is likely to continue to drive in the marketplace," said Intel's CEO Paul Otellini, speaking to the reporters.

Not so fast, said IBM.

Big Blue heard of Intel's plans to announce the High-K + metal gate transistors. It managed to secure a pre-press version of Intel's news release, which was originally scheduled to hit the news wires on Jan. 29. So, IBM quickly wrote up a news statement of its own and handed it out to reporters under what's known as an "embargo" where reporters promise to release all of their news stories at the same time. (The "embargo" is an archaic technique used by vendors to control coverage of their announcements and get the most possible press at one time. It also gives print publications a chance to include fresher stories in their pages by getting "news" ahead of time. Most reporters hate embargoes but always agree to them.)

You can tell that IBM was under serious time and creativity constraints with its news release because its language is not all that fresh.

For example, Intel's press statement notes that the transistor breakthrough "also ensures Moore's Law, a high-tech industry axiom that transistor counts double about every two years, thrives well into the next decade."

IBM glommed right onto that fancy 'axiom' talk.

"As a result, the use of this material could allow the industry to continue on the path defined by 'Moore's Law,' the chip industry axiom that predicts a doubling of the number of transistors on a chip every 12-18 months, thereby allowing chip performance and function to increase as well."

Intel's pushed out its news release under the headline - "Intel's Transistor Technology Breakthrough Represents Biggest Change to Computer Chips in 40 Years."

IBM went with "First fundamental change to basic transistor in forty years."

Unfortunately for IBM, the company eventually had to stray from Intel's language by admitting that it's not quite as far along as Intel with the new transistors. IBM would only commit to shipping products with the new transistors "starting in 2008."

Somehow, IBM doesn't see Intel's clear lead as a lead.

IBM's VP Bernard Meyerson told the New York Times "that industry analysts who have asserted that Intel has a technology lead are not accurate and that I.B.M. had simply chosen to deploy its new process in chips that are part of high-performance systems aimed at the high end of the computer industry."

As far as we can tell, that means IBM will use the technique first with its own Power chips or with AMD's Opteron chips. So, IBM has either screwed partners/customers AMD, Toshiba and Sony by giving preference to its own chips or is in fact well behind Intel's 45nm Xeon chips.

In its rather short news statement, IBM declared that it would share more information about its High-K + metal gate transistor breakthrough by "publish(ing) the summary of this final achievement" at an industry conference sometime this year.

Well, wouldn't you know it, we just happen to have acquired a rough version of that very presentation. Geeks out there can read up on IBM's breakthrough ahead of time via this PDF - a Register exclusive.

Doesn't it suck when someone messes with your timing? ®