March 13th, 2009
The XO-2 is going Embedded RISC? So will everyone else!
Yesterday, OLPC project founder Nicholas Negroponte indicated that their next generation XO-2 will almost certainly eschew it’s current x86-based design and move towards one that is based on the ARM, the most common embedded RISC architecture in the world. It’s used in over a billion cell phones and smartphones, including the iPhone and the BlackBerry, and is licensed for use in embedded CPUs produced by Texas Instruments, Marvell, Broadcom, nVidia, Samsung, and FreeScale, among others.
Click on the “Read the rest of this entry” link below for more.
Now, I know what many of you are probably going to say — OLPC isn’t exactly the most successful project in the world, and on the grand scheme of things, they’ve had minimal impact on the industry in terms of their ability to ship units. However, the design goals of their devices have very much influenced what has been happening with mobile computing as a overall trend, such as the current netbook revolution. OLPC’s fumbles aside, nobody can deny that low power, long battery life and lower cost portables are the future. And the only way that future can be realized is if x86 on mobile systems goes away.
I’ve been predicting that the x86 was going to disappear for a while, but it never occured to me that the economy was going to be the driving force behind it.
I think it is probably safe to say that the next generation of the Atom chipset, Diamondville, is almost about about as far as the x86 architecture can progress in terms of its ability to conserve wattage and also provide the same balance in CPU horsepower needed for a MID or a general purpose mobile computing device. Right now, for your typical netbook, the total Atom power consumption of CPU + support chips is around 11.8 watts, with the CPU itself consuming about 5 watts. Diamondville’s support chipset will consume about 2.5 watts, bringing the total to around 5.5 to 7.5 with CPU depending on whether you use a single core or a dual core Atom chip. And While the Atom support chipset provides for graphics control and sound, that doesn’t include display, wireless, storage or anything else that is needed to provide a complete mobile platform– which consume more power.
The reason why x86 cannot progress any further for mobile systems is because of the sheer amount of transistors that have to be put on a die in order to support all the legacy instruction sets for x86 compatibility. With RISC processors, no such compatibility is needed, so less transistors are required and less power is consumed. It also means you don’t have to clock the chip as high to achieve good performance with embedded applications.
OLPC has seen the writing on the wall — RISC. With chips such as the TI OMAP that power BeagleBoard-like devices (which can have complete platforms operate in 2.5-3 watts of power) or the Marvell XScale which also use ARM-based designs, the requirements for power efficiency, manufacturing cost and chipset integration will override the need to use the legacy software base. That being said, new OSes will have to be used on these systems instead of what we use on most of today’s destkop computers.
There are two OSes which are well suited to these low-power chips — Linux (including Google’s Android) and Symbian, both of which are Open Source. iPhone OS and Windows Mobile/Windows CE also can be used, but both are proprietary. iPhone OS will never see the light of day on anything other than a device made by Apple, thus limiting the potential market size because whatever Apple makes will never be cheaper than the alternatives, and Windows CE/Windows Mobile is going to need a lot of work for it to become a robust mobile computing platform which would provide near-desktop experiences on a MID.
Certainly, a full-blown Windows port to ARM is not out of the question, but it would be a significant undertaking. Whether Microsoft has fully considered putting that much resources into mobile beyond what it has currently committed for Windows Mobile 7 remains to be seen.
Will the mobile future be a RISC-based one? Talk Back and Let Me Know.
Loading ...![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=fc7b52d3-fe5f-45d0-a39d-caf171632a08)
Jason Perlow is a technologist with over two decades of experience integrating large heterogeneous multi-vendor computing environments in Fortune 500 companies. See his full profile and disclosure of his industry affiliations.
Subscribe to Tech Broiler via Email alerts or RSS.
| 03/13/09
