With many-core processing leveling the playing field for microprocessor makers this would be a good time for AMD to focus on system-level issues impacting its processor ICs and let ARM do the heavy lifting in terms of architecture, argues Peter Clarke
It thus becomes possible to read some significance into the presence of Microsoft and ARM executives as keynoters at an upcoming Fusion development conference. The event will run June 13 to 16 in Bellevue, Washington, which is Microsoft's backyard, and include more than 50 technical sessions spanning areas including multimedia, user interfaces, business and high-performance computing and security.
The event might provide a good opportunity to announce the extension of Fusion to accommodate licensed cores such as Cortex-A series cores or Mali graphics along-side x86 cores.
A board-level strategic decision to go with ARM might also fit with the departure of AMD's CEO. When Dirk Meyer quit as AMD CEO in January 2011, reports suggested he was forced out by a board of directors dissatisfied with AMD's lack of chips for the tablet computer and mobile market.
Now it may be that a hard-pressed AMD was simply unable to create designs for all the different sectors and performance profiles in a proliferating PC landscape. But that alone is reason to get off the x86 treadmill and let ARM do some of the heavy lifting. And having missed the tablet computer boom ARM licensing would provide the fastest way for AMD to get a chip to market and make up lost ground.
A licensing move by AMD would, of course, exactly fit with ARM's commercial battle with Intel. Indeed, it would recognize ARM taking over from AMD as Intel competitor-in-chief as well as pushing back on the global chip leader painting the company into a tighter corner.
If it came to the point where the x86 architecture is not worth second-sourcing what does that say about the value of the first-source chips?