According to an SEC filing, he quit last Friday after 20 years and assorted jobs at Sun.
He’s only been running the Sparc operation for a year. He was supposedly unhappy running storage, his previous job, and thinking about leaving Sun then so he was repotted.
His immediate replacement – until they find a permanent one – is Mike Splain, CTO of Sun’s systems group and a Sun Fellow.
Sun hasn’t had many takers on its attempt to get other people using the Sparc by open sourcing the spec. It’s recently lost its hereditary Texas Instruments processor fab and will have to shift 65nm Sparc production to TSMC. It’s also delayed the arrival of the so-called Rock processor ostensibly to give Niagara more time to establish traction in the market.
Ironically on Monday Sun said it’s getting $44.3 million in funding from the Defense Advanced Research Projects Agency (DARPA) to advance its Proximity Interconnect study of chip connectivity using laser beams instead of wires between the chips to move data faster, especially when thousands of chips are required to crack a difficult problem.
Sun beat out Intel, HP and IBM for the privilege apparently because of its alignment techniques – the chips have to align just right.
The New York Times says that if Sun’s research pans out it could mean “more compact machines that are a thousand times faster than today’s computers. Each chip would be able to communicate directly with every other chip in the array via a beam of laser light that could carry tens of billions of bits of data a second.”
Very Flash Gordon and a very “high-risk program,” according to Sun, which figures there’s a 50% chance of failure. It’ll take it five, five-and-a-half years to find out whether what it calls a “macrochip” – because the wafers would be way bigger than today – has commercial viability.
Sun will be working with Stanford and the University of California at San Diego as well as two silicon photonics firms, Luxtera and Kotura.