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
LONDON – Warren East, CEO of processor licensor ARM, says there is no story here.
East says that he and his CFO Tim Score were asked about AMD during analyst discussions over ARM's record first quarter financial results and that he had simply said that ARM has been trying to sell to AMD for long time – as ARM's shareholders would expect. And as AMD is rethinking its strategic options there is clearly a "heightened opportunity" to make that sale.
That AMD might abandon – or at least augment with licensed-in ARM processors – the x86 processor architecture that has defined the company for 20 years, is to think the unthinkable. But it clearly makes sense. Such a move could not happen overnight. It may take years with x86 processors from AMD hanging on in some applications but the general arguments seem compelling.
And such a move supports ARM's increasing rivalry with Intel and its position in the IBM/Globalfoundries Common Platform Architecture camp.
Since AMD made providing x86 cores its primary business, it has always been in the shadow of Intel. This is partly because as Intel is the definer of the architecture, AMD is always playing catch-up; trying to provide a code-compatible processor at a better price. And Intel, as market leader with phenomenally deep pockets, has been able to use a mixture of pricing and manufacturing leadership to keep AMD under pressure.
Indeed, AMD was forced to divest itself of in-house manufacturing two years ago as it could no longer afford to keep developing manufacturing technology and putting down billions of dollars to build wafer fabs in which to run them at the same time as developing microprocessors. That divestiture was the seed for the creation of Globalfoundries Inc., now one of its foundry partners.
Most users can really save with AMD chips instead of Intel's because they dont really need Intel chips to do the jobs most of the times. IT teams often picked Intel chips just to play it safe but AMD chips can really do the jobs for less. Resellers seek to maximize revenues with higher priced Intel chips inside the throats of users.
Maybe there is a need for benchmark software to help IT guys analyze their own wares to determine if AMD chips can do the jobs every bit as good as Intel's. AMD enjoys a very low market share in the server business to make me suspect a whole lot!! at 7% of the server market just do not jive !
Another possibility is for AMD to support both architectures. They can support the x86 for markets such as PCs; and ARM for mobile computing and smartphones. Or augment their x86 product with an ARM processor; but that could turn out to be an expensive solution.
For ARM this is a good move, because AMD would add another licensee to their ecosystem; moreover it can help them get into computing markets where they traditionally haven’t had much traction. For AMD it makes sense because its destiny will be defined by more than what its chief competitor (Intel) does. The only caveat for AMD would be that, they would now enter a chip space with a sea of competition including Broadcom, Freescale, Infineon, Qualcomm, Samsung, ST Micro, and TI.
I think this is a great strategy idea for AMD. They will not win a head-on competition with Intel, yet they have tremendous resources and skills at their disposal. Qualcomm has done a nice job with their Arm-Licensed Snapdragon processor...optimized for mobile use. AMD could choose another tangent such as chipsets for low-cost, low-power multiprocessor servers (like SeaMicro except using ARM's instead of ATOM's).
ARM based processors lend themselves to the low current requirements for smartphones, tablets, and small laptops. In this sense, AMD might benefit by being able to sell into this rapidly expanding market. However, there are lots of competitors with their own ARM based processors, including Apple, Qualcomm, Nvidia, and Samsung. Qualcomm has years and years of experience combining its Snapdragon (ARM based processor) with its modem chips on a single chip, with smaller size, lower current draw, and other features that offer challenges to less experienced competitors. What can AMD offer to survive in this market?
This is a bad deal for AMD but good for ARM and Intel. AMD will be joining a long line of ARM licensee's all trying to out do one another with basically the same arch (ARM). AMD is currently in a market where it hold #2 (very distant but still #2) and has lots of experience at x86. They will be entering a crowed market which they have no expertice. Intel will pretty much own 100% of the x86 market at that point. ARM just gets sell yet another license.
AMD do not have to drop x86, why should they? That said, I think such deal will be more of an advantage to ARM than AMD. As for Intel, I would be very worried if I were in their shoes. I do not see increased x86 dominance as a blessing, certainly not in the mobile computing era.
I'm wondering why Nvidia is off the radar from this analysis. Its Project Denver is arguably much further down the planning and development path than anything AMD could start doing even yesterday. Also, Nvidia is claiming the entire high/low TDP range for Project Denver. Not only would it make it a much more ambitious project than what AMD could conceivably reconsider with ARM right now but it would be more complete as Nvidia already have an ARM-based portfolio.
The possibility of only one probable outcome is highly likely; AMD will most likely license ARM cpus. 2 factors help the domination of X86 programs- the amount of installed X86 software, and shrinking die size of future Fusion APU's, Intel has gotten a 30% increase of performance and battery life with every die shrink - Llano should gain, as well as the C class APU's - Monte Carlo simulations and simultaneous HD at the same time, on a tablet or netbook?
January 2016 Cartoon Caption ContestBob's punishment for missing his deadline was to be tied to his chair tantalizingly close to a disconnected cable, with one hand superglued to his desk and another to his chin, while the pages from his wall calendar were slowly torn away.122 comments