LONDON – Processor licensor ARM Holdings plc is trying to persuade Advanced Micro Devices Inc., the long-time rival to global chip leader Intel Corp. to license ARM processors and use them instead of the x86 architecture.
It is notable that software giant Microsoft and ARM (Cambridge, England) are due to provide keynoters for an upcoming Fusion development conference and that Microsoft recently announced that it would support its full Windows operating system running on ARM processors.
One possibility is that AMD could amend its Fusion architecture to include both x86 and ARM CPU cores plus graphics cores in a heterogeneous multiprocessor. However, it is also possible that AMD may choose to abandon difficult and expensive attempts to develop its own multicore x86 architecture in competition with Intel and, instead, allow ARM to provide its cores.
Speaking to EE Times during a discussion of ARM's first quarter financial results CEO Warren East said: "AMD is a successful company selling microprocessors. ARM is in the business of licensing microprocessor designs. It is perfectly natural that we should have been trying to sell microprocessor designs to AMD for about the last ten years. Hitherto we haven't been successful."
East also said: "AMD has signaled they are going through a rethink of their strategy, and that must provide a heightened opportunity for ARM. They might use ARM microprocessors in the future and you've got to expect that we would be trying to persuade them of that."
"One possibility is that AMD could amend its Fusion architecture to include both x86 and ARM CPU cores plus graphics cores in a heterogeneous multiprocessor."
hardly, AMD or rather the ATI executives that now control much of the executive board now, could/did not even see fit to make the very latest Fusion "Ontario" architecture product that's been in development for 5 years have more than SSE3 SIMD on board.
apparently its got the old SSE/SSE2/SSE3 extension's but no SSE4a and no AVX.
look at all the SIMD x264 benchmarks over the years and you see that AMD have never once beaten Intel in these hand written assembly SIMD scores clock for clock.
so what hope is there that any AMD executive could make a viable choice to make hybrid x86/Arm Fusion architecture that can provide something better clock for clock than the other vendors in the future.
Drop x86 and its revenue stream? That's suicide. Add ARM for mobile? Sure, if they can afford to increase their R&D expenses and work on ARM-based processors in addition to what they're already doing.
A low-power version of an ATI GPU plus a pair of ARM cores could be an interesting product...sort of a Fusion-lite for tablets.
I can't see it being a smart business decision to leave a market that has essentially one competitor to be a late-comer in a market that has (presumably) lower margins and a load of competitors. Adding in a low-watt processors to get into the tablet and phone markets might make sense, but as others have stated, not at the expense of AMD's core market.
AMD is already fabless. If they also go to IPless, where is their added value? I really like ARM, their products and their business model. I think they do a great service toward ubiquitous computing, but this to me, reads like either a publicity stunt or some misguided arrogance.
There's two sides to all these coins. Right now AMD is number two in a huge x86 market that barely has a number 3 player (Via). X86 PCs running Windows and Linux ain't going away in my lifetime--such as it is. In the ARM world, AMD would have to compete with Freescale, TI, Qualcomm, Marvell, Nvidia and others all piling on the smartphone and tablet markets and tryoing toi edge into servers where they lack software support. ODMs already say there isn't much differentiation between these companies. Good luck to AMD being tenth in those markets.
I agree it does not make sense for AMD to drop x86 immediately. No doubt there is roadmap of designs coming through that will impact the market for some time.
BUT sticking with x86 is costly in terms of engineering resources while licensing ARM cores would be economically efficient.
And once both approaches are in-house, with Windows running on both, it would allow direct comparison.