IDF Dabblings
Two weeks ago, I spent Tuesday and Wednesday at Intel's twice-yearly US Developer's Forum. By now, lots of cyber-ink has been spilled regarding the goings-on, so I'm not going to bother regurgitating the raw information you can already find at top-notch sites such as AnandTech, ExtremeTech, The Register, Slashdot and Tom's Hardware Guide. You can even pretend you're a tech journalist too (don't worry, the transformation isn't permanent) and visit Intel's press site for the releases, keynote transcripts and other goodies. And according to Intel's IDF website, the spring 2005 conference presentations will be available to non-attendees in May (although if you visit the fall 2004 conference presentation cache you might find an interesting login/password combo listed…..wink wink, nudge nudge, know what I mean?).
What I do want to do is blog my take on things that I personally saw and heard (and found particularly intriguing). Big-picture, the spring 2005 IDF was a return to past form that I hadn't experienced from the company and its partners since my days as an Intel employee and booth bimbo, and maybe the first year I subsequently attended as a press hound. The overall mood was calm, subdued and very confident, and although AMD bankrolled an airplane to fly above Moscone Center and script Turion 64 (nothing like rebranding an already-announced 90nm Mobile Athlon 64, eh?) in the sky with smoke, I gotta believe Hector Ruiz and the rest of the folks in Sunnyvale are quaking in their boots now that the Intel super-tanker has completed its slow turn and is headed full steam in AMD's direction.
So what happened between 1997 and now? The first few intervening years, my perception is that Intel got all caught up in the Internet boom and the tumult of all the companies and technologies it gobbled up. Presentations became a blurred morass of visions in dozens of market segments, and no amount of hand-waving and spin could tie them into a coherent, cohesive strategy. Also, Intel got greedy and tried to tighten its strangehold on the PC architecture (and industry) via initiatives like the Rambus experiment.
About the time of the dotcom collapse, the cracks in Intel's frenzied plan became too evident to dismiss any longer. RDRAM was a market failure, as were many of the acquisitions the company had made. AMD's Athlon CPUs were faster and lower-power than the Intel counterparts, as evidenced by their embrace by the hyper-critical gaming community, and AMD's clock speed and 64-bit leadership put Intel on the defensive for the first time in many years. Intel's presentations inherited an edge of desperation and, while the company frantically reoriented itself behind the scenes, IDF for several recent iterations was, in comparison to shows of years past, a no-new-news event.
Those past times of turgidity and turmoil appear to be behind Intel now, though. Yes, there were the obligatory mentions of Xscale for advanced cell phones, WiMAX, flash memory, and other peripheral products and technologies, but Intel's primary focus at the Spring 2005 IDF was once again on the PC. Dual-core CPU announcements were everywhere, of course. Intel's Stephen Smith gave a very revealing presentation on the company's plans through the remainder of the decade, across all computing segments, and the company provided an accompanying 'decoder ring' fact sheet. Believe it or not, although Intel's provided me with a PDF of the presentation and a DOC of the fact sheet, their lawyers insist I can't post the files for you to download. What I can do, though, is post screenshots. OK…..(insert obligatory rant about large corporation bureaucracy) …..anyway, see the links to JPGs below:
CPU roadmap (which doesn't include Intel's prediction delivered at Corporate Technology Group Director Justin Rattner's Thursday morning keynote that, by 2015, the company would be able to integrate 100 cores on a single chip)
Forecasted dual-core market penetration by end of 2006, across desktop, server and mobile segments
Multi-core implementation options
Pentium (note no '4') Extreme Edition 840 details
Smithfield aka Pentium D (ditto on '4') details
Presler details
Cedar Mill details
Yonah details
955x Express core logic chipset details
Fact sheet pages 1 and 2
Fact sheet pages 3 and 4
For more on the Spring 2005 Intel Developer Forum, see my next blog post, IDF Double-Dabble
