Microsoft Signs License With ARM

G143 and several other readers let us know that Microsoft has signed a licensing deal with ARM. "Microsoft signed an agreement with the UK-based ARM, giving Microsoft access to some of the chip designer's intellectual property. The two companies have worked together since 1997, but Ian Drew, ARM's EVP of marketing, said this is the first time Microsoft has become a licensee of ARM's architecture, a move which will allow Microsoft to design their own microarchitecture. Other licensees include Qualcomm, Marvell, and Infineon. Neither company would reveal the cost of the license. Speculation about Microsoft's intentions includes wondering whether the company is taking aim at the iPad, or perhaps looking to produce a next-generation Xbox without the 360's heat problems."

Already done, thank you very much

[garyisabusyguy]

The ARM core is so widely licensed that it would be hard to find a modern handheld device that does NOT contain one.

"Many semiconductor or IC design firms hold ARM licenses; Analog Devices, Atmel, Broadcom, Cirrus Logic, Energy Micro, Faraday Technology, Freescale, Fujitsu, Intel (through its settlement with Digital Equipment Corporation), IBM, Infineon Technologies, Nintendo, NXP Semiconductors, OKI, Qualcomm, Samsung, Sharp, STMicroelectronics, Texas Instruments and VLSI are some of the many companies who have licensed the ARM in one form or another" http://en.wikipedia.org/wiki/ARM_architecture#ARM_licensees [wikipedia.org]

IMHO, this is a non-story.

Re:Already done, thank you very much

[garyisabusyguy]

I had a couple of 'why' questions and found a possible answer.

This Intel forum:
http://software.intel.com/en-us/forums/showthread.php?t=67843 [intel.com]

Contains many complaints about the performance of the Intel compiler under WinCE (as well as Intel selling it all off to Marvel).

If you ask me, MS wants a chip that they can optimize for their OS. Seems liek this will lead it down a proprietary hole, not unlike Apple.

So, is that the real story, "MicroSoft, now more like Apple!"

Re:Hmm...

[Amouth]

i'm betting on 5 things - 1-5 most likely to doubt

#1) and and most likely (as it has the highest chance to fail) - MS tries to make it's own hardware for it's Win7 phone OS so that they don't get the bad rap they did with windows mobile on phones that didn't have the hardware to run them

#2) they might be looking into using it for the next xBox - and at that point doing it in-house so they don't have to rely on IBM as they have in the past (them supporting IBM who is also being supported by Sony can make it seem like they are supporting the competition)

#3) they might be looking into using them for their next HPC platform - it is at least 3-5 years out which is a good lead time for them to design and refine a new way of using exiting ARM cores.

4#) maybe they want to design and test extensions to the ARM archt that they don't want to trust a partner with - once they refine them submit them back to ARM (think of it as custom extensions for either Win7/8 Phone OS or xbox OS or HPC OS)

#5) maybe they are rethinking their canning of the courier - and are rather going to embrace it and actually make it with a competitive chance.

I find it odd that in my mind the list 1-5 of most likely to doubt is inverse of what i think would work out best for them.. and would be happy to see them do. Maybe its All of the above and a slice of pie?

Windows phone

[Local ID10T]

Speculation about Microsoft's intentions includes wondering whether the company is taking aim at the iPad, or perhaps looking to produce a next-generation Xbox without the 360's heat problems.

Seriously? Microsoft has been chasing the smartphone market for a while now, but keeps having performance issues. They want a custom designed chip for their next gen smartphone.

Re:XBox Portable?

[ravyne]

Thats because ARM cores have always been targetted at the embedded and portable markets, where potential customers want "just enough" power with absolute minimum power draw -- Tha majority of ARM parts sold today draw peak power measured in milliwatts -- an order of magnitude or two lower than the most frugal X86 or PPC designs. Some of the higher-end ARM designs incorporate vector floating point, a DSP or two, hundreds of megabytes of on-chip RAM and quite capable 3D accelleration, and still manage to draw down less than a watt at peak power consumption.

The fastest available ARM cores today are competitive with Intel's ATOM processors core-for-core, clock-for-clock. In practice, existing cores end up being slightly slower in actual use since none have a system bus wider than an anemic 32bits (and also not typically clocked very fast), but in terms of computational power, they already match some current x86 designs.

If someone took the chains of power-consumption off and said "We're going to get all we can out of, say, a 20 Watt power envelope" you'd end up with something quite powerful. For starters, with very little work on the core itself, you could double the core speed and simply allow greater power leakage, then toss 16 of those on a die with local and shared cache. 16 2Ghz ARM cores, each with a very capable vector-floating-point unit, would be every bit as powerful as even the fastest Intel quad-core i7. Then there's the other end of the spectrum -- reworking the core architecture from the ground up, optimizing for performance, rather than power consumption. You could reasonably attain competetive performance with fewer, more-powerful cores as well.

All of that said, I'm skeptical that this is about Xbox. This is for sure about Zune, possibly about a Microsoft portable console, possibly for high-performance peripheral devices (Think Kinect 2.0 offloading most of the work to the kinect unit itself), and only just by the slimmest of chances about Xbox 3 (or 4, for that matter) -- The only thing that gives me pause that it might be about Xbox proper, is that all their other devices are well-served by the existing ARM ecosystem, too well served to enter the chip-design business anyhow. An upcoming Xbox is one compelling reason to liscense a new micro-architecture, and being able to respond to Apple's in-house ARM core is the other.

Make no mistake that ARM is a very capable architecture though, and it's a delight to code for in Assembly Language. I've said many times that ARM is, in the coming years, going to be the most credible threat Intel and X86 has ever faced. It's going to be moving into Netbooks and STBs more this year, then laptops, some servers and higher-end embedded devices (I'm convinced Nintendo will consolidate all their systems behind ARM beginning with their next home console in order to leverage their 1st and 3rd party experience with ARM that is the result of their portables' popularity). In 10 years time, the question may not be Intel vs. AMD, but x86/x64 vs. ARM.