Sun Sparc guru splits for Redmond
When it rains...
In stunning blow to Sun Microsystems, the company's lead chip designer, has resigned. And Marc Tremblay is reportedly taking a job at Microsoft.
Sun has not been particularly talkative about anything since the rumor of IBM's proposed acquisition of Sun broke three weeks ago, but a company spokesperson offered this official statement on Tremblay's departure:
"Sun can confirm that Marc Tremblay, Sun Fellow and CTO for Microelectronics, has decided to leave the company. Rick Hetherington, who has served as co-CTO for Microelectronics with Tremblay, will assume leadership. We thank Marc for his many contributions over the last eighteen years and wish him all the best in his future endeavors. Hetherington has been with Sun for more than ten years and has served as co-CTO for the Microelectronics business unit for two years. Prior to that he spent sixteen years with Digital Equipment Corp. working on various processors and systems."
Sun's chip business has been problematic since the turn of the millennium, and Tremblay could be leaving because he had enough or because of delays in bringing the "Rock" UltraSparc-RK processors to market, which are expected some time in the second half of this year after being pushed out for approximately a year.
Back in March 2008, David Yen, Sun's former top chip techie and executive vice president in charge of Microelectronics, left the company to take a job at Juniper Networks. Sun has tapped Mike Splain, a technology fellow and the senior vice president and chief technology officer of its Systems Group, to be the acting head of the Microelectronics group.
This was just after the Rock chips were pushed out, and at the time, tongues were a-wagging that Yen was pushed out of Sun rather than jumping to a new job as executive vice president of emerging technology at Juniper. In November 2008, when Sun announced up to 6,000 employees were to be laid off, the company also reorganized itself, bringing chips, servers, storage, operating systems, virtualization software, and related systems management programs all together into a single Systems Group, under general manager John Fowler.
A month later, when Sun co-founder and CTO for the Systems Group, Andy Bechtolsheim, left Sun to become chairman and chief development officer at a startup called Arista Networks, Sun tried to spin it and say that he was still at the company and still chief technology officer for the system biz. Not so with Tremblay, who is an equally important techie in Sun's army of engineers.
If you want to read Sun's official bio of Tremblay, you can't because the company has whited it out on the executive biography section of its Web site. But Tremblay was the senior vice president and CTO for Microelectronics and also a Sun fellow. He was the co-architect for the UltraSparc-I "Spitfire" processor designed by Sun in the mid-1990s and that gave it credibility in the Unix server space Sun had been making various Sparc processors since 1985, but the UltraSparc-I gave Sun a big leap in performance.
Tremblay was more importantly the chief architect of the "Blackbird" UltraSparc-II processor, announced in 1997, that powered Sun's incredible growth during the dot-com boom. Tremblay was not in charge of the UltraSparc-III designs (which were less than Sun had hoped), Yen pulled the UltraSparc-IVs out of the fire, and Tremblay has been in charge of the design of the forthcoming Rock chips. Not everything Tremblay did turned into gold: the MAJC Java processor, which was busted down to a silly graphics card, and the picoJava processor for running Java bytecodes, were failures.
His departure from Sun last week, for a job at Microsoft according to a report in the New York Times, does not inspire confidence in the future Sparc chips, even though chips are designed by armies of people, not one person. But you need really smart people to lead those armies in the right direction as they forge the weapons with which vendors fight their battles for the budget dollars in the IT market.
What Microsoft needs a Sparc chip geek for is another good question. But it is clearly better to take Tremblay out of the field and give him money and intellectual challenges for x64 processors than to have him give the Sparc platform his best ideas. ®