Earlier this week, we reported on rumors that AMD’s Zen might have slipped into Q4 2016. Since then, we’ve heardthe chip could actually launch in the Q1 2017 timeframe — and now, there’s further reason to think that something happened to AMD’s next-generation CPU timetable. Now, it’s been reported that Jim Keller, who returned to AMD to helm its new CPU after a stint with Apple, has left the company to “pursue other opportunities.”
Keller wasn’t just responsible for the Zen CPU architecture; he was also leading the team that designed AMD’s still-upcoming ARM-based K12 CPU, which isn’t expected to launch until 2017. AMD sought to downplay the impact of this announcement and told Hexus.net that “Jim’s departure is not expected to impact our public product or technology roadmaps, and we remain on track for “Zen” sampling in 2016 with first full year of revenue in 2017.” Mark Papermaster will now step in and head Keller’s team.
An uncertain impact
The knee-jerk way to read this announcement is that Jim Keller was fired because Zen is coming in 6-7 months late. That’s entirely possible, and it wouldn’t be the first time an AMD executive left to pursue “other opportunities” for reasons that only became clear months after they were gone. Dirk Meyer’s departure as CEO didn’t make much sense at the time, and it was widely reported that he was forced out over disagreements related to the tablet and mobile markets. Later, once Bulldozer had hit store shelves, it became clear that tablets and mobile products hadn’t been the only problem.
AMD’s Zen CPU
When you consider the differences between the AMD that Keller came back to in 2012 and the AMD he left in 2015, there’s no shortage of factors that might have caused a break-up. In 2012, AMD was clearly planning to enter the ARM market and launch its own custom ARM core (and Keller’s most recent expertise was in ARM SoCs, not x86 processors). In 2015, the K12 and Cortex-A57 CPUs that Sunnyvale once championed scarcely warrant a mention.
Skybridge went from mission-critical to dust in less than a year
As recently as 2014, AMD had a public roadmap for a common socket platform between x86 and ARM cores that would bridge the two, with an HSA-enabled version of the Jaguar architecture that might have helped plug the holes in AMD’s roadmap between now and Zen’s launch in 2017. By 2015, those plans had been canceled. The recent graphics reorganizationand rumors of substantial private equity investments could be further indications that AMD’s new focus isn’t what Keller signed on to shepherd, and that he’s decided to pursue other opportunities without it being evidence of a substantial problem in AMD’s product pipeline.
One thing we’ve heard from multiple knowledgeable sources is that Zen is finalized. We don’t know if the architecture has taped out or not, but at least the vast majority of the work is already complete. I’m reminded of a quote attributed to Robert Palmer, the ex-CEO of Digital. “Designing microprocessors is like playing Russian roulette. You put a gun to your head, pull the trigger, and find out four years later if you blew your brains out.”
Keller’s departure will not be well-received. Here’s hoping it had more to do with differences over the company’s focus as opposed to Zen itself. AMD is out of time for putting its own house in order.
No comments:
Post a Comment