And certainly setting up a foundation is the relatively easy part. It is
easy compared making something that is designable, energy efficient and
EASY TO PROGRAM that covers square centimeters of silicon with billions
of transistors and built using 14- or 11-nm design rules.
But Intel's success in the personal computer forced AMD forced to take the long view.
is missing are some examples of success, such as an Apple, Nvidia or
ST-Ericsson HSA engine. That may have to wait on deployment of the
PowerVR Series 6 and the Mali-T6XX graphics cores which are designed to
support more general computing as well as graphics rendering.
should go to AMD's CEO Rory Read who was prepared to roll the dice on
AMD's future, but in their previous state AMD had very little to lose,
except more money. Kudos should also go to the senior executives at ARM and Imagination who were able to set aside graphics rivalries to be prepared to work together within HSA. Will Intel be able to do the same?
Now the difficult business of writing software
and defining software interfaces resumes. It is to be hoped that more
companies will quickly join this initiative. It will need all the brain
engineering the industry can muster, but this looks to be a good start and pointed in the right direction.
Related links and articles:
Analysis: Why ARM-AMD makes sense
AMD to integrate ARM core into APUs
AMD, ARM, Imagination, TI, MediaTek form HSA group
AMD creates embedded IC business unit