IBM Eye

I’ve been pretty slack over the past year with rumors, mostly because I’m not sure how to confirm them. I dislike whisper mills so I’ve shied away from them unless I feel pretty good about what I’ve heard. But this is, after all, a blog, so I feel like I can take some liberties. On that note, I’ve heard a couple things that I don’t have the contacts to confirm so I figured I’d just throw them out there.

1) Sometime next week (the 24th) there will be another round of corporate wide layoffs for big blue. This is fairly commonplace… so much so that it may not really be big news from a corporate perspective, except that I’ve heard that in New York, IBM has arranged extra security to escort those dismissed from the premises. It is my understanding that that level of ‘protection’ wasn’t taken for any of the other recent layoffs.

2) IBM Research chief Paul Horn retired today after 28 years at IBM (11 as head of research) I was tipped that he was pushed out and it wasn’t amicable. I don’t know why or anything, but I thought other might clarify things in the comments.

IBM has been in the news this week for announcing a breakthrough in High-k Metal Gate Technology. Though IBM was the lead, they partnered with AMD and others to share risk and expense. Intel singularly announced a similar breakthrough the same week. I’ve provided what I believe is a sufficiently detailed summary of the breakthrough (I’ve been summarily disappointed by every article I’ve read). If you’re interested in the physics or math of it all, Wikipedia is, as always, fantastic.

Background:

Moore’s law for processor speeds is in jeopardy because of limitations imposed by the use of Silicone Dioxide to insulate circuitry in microchips. One problem with SiO2 is that 5 atoms thick is as thin as they can get it, imposing a hard limit on how small the chips can get. The second problem is that at that width, noticeable current begins to seep out (technically “leakage currents due to tunneling” results), lots of heat builds up and lots of power is consumed. It has proven very difficult to find a suitable replacement material.

The Solution:

IBM et al found that certain hafnium http://en.wikipedia.org/wiki/Hafnium alloys (probably Hafnium silicate - HfSiON) can be used as a more effective insulator than silicon dioxide and is planning to use the element to produce faster and more energy efficient chips, by allowing circuitry scaling to be reduced to less than 45 nanometers. Integrating high-k / metal gates will address the power consumption issue; a major barrier to scaling chips and continuing with Moore’s Law. It is thought that with this weeks announcements, Moore’s law has been extended through the next decade.

Another huge benefit to the Hafnium Solution is that it can be implemented without requiring major tooling or process changes in manufacturing. It seems that most of the alternative gate dielectric materials were impractical for existing manufacturing facilities resulting in potentially Billions of Dollars to move to the next generation of chip circuitry.

On-Topic Tangent:
Dr. Rajarao “Raj” Jammy was the project lead on this. He has 50+ patents and is one of those guys that you only ever see working in a University somewhere or at IBM. Anyway, there’s a great interview with here at Reed Electronics following a conference on High-k dielectrics. A quote:

…(In) reference to SiO2 or silicon oxide or silicon oxynitride gates, those gates with polysilicon electrodes have stopped scaling. So once these dielectric stacks or gate stacks have stopped scaling, we really had no option left. But for performance enhancement, people continue to use new ideas, like mobility enhancement. And therefore, they continue the scaling that the industry is so used to. But in some point in time we have to get back to the dielectric and try to see how we can continue to scale the dielectric also – part of it for the improvement in the coupling that we achieve at the channel, and therefore drive more current; but also the key part of it is reducing the leakage that comes from the gate dielectric.

The New York Times had a front page article this week talking about IBM’s announcement that they are trying to add more transparency to their patent filing process. In summary they will:

• Put all their patents on the web whether current or pending
• Try to name their patents intuitively
• No patents based on business processes unless there is substanative technical merit

IBM’s release can be found here.

There’s a lot of great discssion on this in the blogosphere. It seems to be overwhelmingly positive, as one might expect from that (this?) group.

The pro:

Feld.com:

This is a huge positive step by IBM and shows real leadership in addressing our completely screwed up patent system.

Woodrow:

Hats off to Big Blue on a very progressive idea, and one that is not without business risk. Ultimately IBM executives have the right idea, which is to say if pending patents are truly innovative and based on technological advances, the fact they’re public poses no real threat.

Techdirt.:
What Does It Say When The World’s Largest Patent Holder Sees Problems With The System?

N/A: Ben Laurie: “I think this is an excellent step in the right direction by IBM”

Seeking Alpha:
IBM’s Push in Patent
Reform is a Very Big Deal

The con:

At a blog called PatentHawk, the author wrote:

To buy the rationale put forth in the NY Times article, that IBM is self-imposing some kind of patent reform, is pure theatre. The move is clearly a publicity stunt, with likely a few unstated motives.

At Jim Moore’s Journal, he bashes “IBM’s Open Source Conspiracy“:

IBM public relations must be singing this morning. They have pulled off “the big lie” at least for now. They have made it look like they are opening up the crown jewels, when what they are really doing is radically diminishing the rights of independent technologists.

It may be that he wrote more to my level, but I found Charles Zedlewski’s rebuttal
to Mr. Moore far more convincing.

For my part, I think it’s a good thing. If a patent is really innovative and
enforceable as it should be, then IBM isn’t giving up much (and gaining PR points and investing in its brand). I do think there’s a lot of posturing here, but the IBM PR engine is often like that.

This news came out a week ago (edit: IBM’s press release came out this afternoon) but I didn’t have anything to add so I didn’t post it. I decided this morning that it was interesting and ought to get a mention just for the news.

Specifically, a supercomputing machine—dubbed “Roadrunner” and set to be fully installed by 2008 at the U.S. Department of Energy’s Los Alamos National Laboratory—will run on some 16,000 Cell Broadband Engine processors and a similar number of AMD Opteron processors. The Cell chip was originally built for Sony’s Playstation 3 console, which has been delayed until November for some markets and until 2007 in others.

IBM and AMD officials said in a press release that the new supercomputer will be capable of a peak performance of more than 1.6 petaflops, or 1.6 quadrillion floating calculations per second. That’s 1600 times faster than the 100 teraflops, or 100 trillion calculations per second, that the average human brain processes. It’s even almost six times faster than the 280-teraflop capability of the world’s fastest computer, IBM’s Blue Gene L

This news is significant for a few reasons. One it speaks to the importance of the work IBM has done with chips for non-traditional uses. By being willing to broaden their horizons, or lose focus, depending on what business school trend is hot, IBM has been able to enhance a core competency.

A second reason this is significant to me is that it demonstrates how IBM can take advantage of advances in one area of its business and monetize it in another. I know of no other company that does as good of a job with monetizing research and patents, and that experience worked to IBM’s advantage here as well.

The third reason I find these sorts of announcements worthwhile, is that it cements IBM in the public mind as being uniquely competent in certain areas of computing. Per a past article in red herring, “Rivals said speed didn’t make IBM’s a better system. ‘We’re much more powerful than IBM’s biggest box,’ Brian Cox, director of worldwide server marketing for HP, told Red Herring”. The guy’s grasping here. If you want someone that really understands fast computing, IBM is who you’d turn to. If you wanted a mid-range PC with Windows XP Media Center edition, you talk to HP. This is an example of why IBM is still relevant in areas other than services and software.

IBM Research has an incredible translation software product dubbed MASTOR (The Multilingual Automatic Speech-to-Speech Translator).  The service has seen rave reviews for both accuracy and ease of use.  At present, it works in both Arabic and Mandarin but the translation engine is very interesting.  Rather than attempting a word by word translation, the engine translates phrases and concepts, known as multi-task and unconstrained dialog translation.  This helps get it past the translation vs. interpreting barrier imposed on all other machine translation.  In brief tests it seems to get the idea across accurately.  The Business 2.0 reporter tested it, saying:

“I ate some bad chicken and my stomach feels bad”

It translated the second “bad” as “that,” but otherwise got it right. Perhaps I was mumbling so I started again, and this time it went without a hitch

Tools like Dragon’s NaturallySpeaking can’t even get the first language right with great accuracy, much less put the words in another language.  This is in spite of the fact that MASTOR requires no tedious and time-consuming training, reading line after line of canned text to help the software learn your voice.  I’m impressed that IBM has been able to pull this off with any sort of accuracy.

As a final note, this project was funded by the Defense Advanced Research Projects Agency (DARPA) CAST program so I’m not sure what if any future windfall from products sales and licensing would belong to IBM.  C-net’s article on MASTOR is entirely focused on the commercial side of the research, but I’m not sure they considered the funding source.

IBM Research has built a history flow analyzing changes over time to a page in Wikipedia.  Something I  think is way more useful and way more educational is this time lapse video of a wiki page.  I get a much better qualitative understanding of how and what changes were made to the page, how long it takes vandalism to be cleaned up, etc. by the video.  Plus the result, in addition to being useful, would have much broader appeal.  A boss would both learn from the video and be wowed by it.  Maybe you can add that in, IBM.