feature article
Subscribe Now

Freescale Bets on “Net-Book” Processor Chip

Just in time for the post-holiday buying season, Freescale has announced its newest embedded microprocessor, the i.MX51. Like other chips in the company’s MX line, the i.MX51 is an ARM-based chip intended for relatively high-end portable systems. Whereas the i.MX31 and ’35 parts are often found inside of automobiles, the ’51 is being touted as the ideal processor for a whole new category of products: the “net-book” computer.

Priced at less than $20 in volume, including its companion power-management and audio-codec chips, the i.MX51 bundle is a bargain. Freescale’s hope is that net-book computer makers will look favorably upon the i.MX51 when comparing it against similar chips from Intel, TI, and Qualcomm. Freescale’s pricing undercuts Intel’s by almost one-half, depending on what you count. Whether that $20 difference in hardware bill of materials will swing the deal is anybody’s guess.

What’s Inside

The i.MX51 is Freescale’s latest design based on the ARM-A8 processor core. If you’re not an ARM aficionado, a bit of a refresher may be in order here. The –A8 is ARM’s current top-of-the-line 32-bit processor design. Following the very successful ARM7, ARM9, and ARM11, the company eschewed simple sequence numbers; henceforth all ARM processors would be called Cortex, followed by a suffix. The Cortex line is subdivided into three ranges: the high-end A series, the midrange R series, and the low-end M series. See how this works? Within each series, there are specific processor designs, such as the M3 or the A8. At present, the A8 and A9 are at the top of ARM’s A-range, and thus represent the current state of the ARM art. Got all that?

In technical terms, the Cortex-A8 is a dual-issue 32-bit RISC processor with a 13-stage pipeline, separate media-processing pipeline, a floating-point unit, branch prediction, and two-level caches (L1 and L2). In other words, it’s nothing at all like the early ARM processors. Nor is it really a RISC processor anymore, for what that distinction is worth. ARM, like virtually all RISC architectures, has quietly adopted all the features it once avoided. In modern processor design, everything old is new again.

Be that as it may, the A8 is small and quick. Freescale has cranked the A8 processor core up to 1 GHz and endowed the i.MX51 with dual 32 KB L1 caches (one for instructions, one for data) and a 256 KB shared L2 cache. These should keep the processor’s dual-issue pipeline well fed at 1 GHz, even though any additional memory will have to be located off-chip.

That memory gets its own on-chip controller, so designers won’t have to construct their own. The i.MX51 handles both standard DDR2 and “mobile” DDR (mDDR) memory chips. Although mDDR uses less power (its raison d’être), standard DDR2 memory is much less expensive. Freescale suggests that customers obsessed with power consumption will opt for mDDR, while those more focused on cost will prefer standard DDR2 memories. It’s nice to know the i.MX51 supports either type equally well.

As if a 1-GHz dual-issue processor weren’t enough, the i.MX51 also comes with two (count ’em) hardware graphics accelerators. One is for OpenGL graphics and one’s for OpenVG video. Both designs are licensed from ATI (now part of AMD) and are intended to offload rendering and decoding tasks from the main processor. Indeed, the A8 processor actually does less work when the chip is decoding video than when it isn’t.

The chip’s peripheral mix reads like a laundry list of standard I/O components: USB, UARTs, IrDA, I2C, SPI, CSPI, 1-Wire, SPDIF, Ethernet, timers, PWM, general-purpose I/O pins, and more. What it doesn’t have is any sort of built-in wireless networking or cellular interface, a seemingly glaring omission in a chip destined for always-on, always-connected “net-books.” And herein lies a little marketing sleight of hand.

The Microprocessor Stork

Although Freescale pitches the i.MX51 as the ideal processor for the emerging (though largely hypothetical) “net-book” market, the chip wasn’t actually designed for that purpose. Product planning and chip design take a long time, and the i.MX51 was well underway before the concept of net-books ever entered the public consciousness. So whether it’s through shrewd product planning or just dumb luck, Freescale managed to design itself a chip that very neatly drops into an emerging market while it’s still emerging. Sometimes even big companies get lucky.

Whatever the background or the reasons, the chip’s wagon is firmly hitched to the net-book star. Only time will tell whether that’s a good strategy. Freescale is positioning the i.MX51 against TI’s dominant OMAP processor family (which combines an ARM processor core with TI’s own DSP) and Qualcomm’s Snapdragon processors (which also combine ARM and DSP). Freescale has also drawn a target on Intel’s new Atom line of low-power x86 processors. This is a bold move, since Freescale (as did Motorola) has a poor track record competing with Intel. And so far, today’s net-books are overwhelmingly x86-based. Freescale has set itself a steeply uphill battle in trying to dislodge Intel once again.  

If the net-book market takes off – and this is far from certain – Freescale might be able to eke out some share of the market. Net-books are, almost by definition, supposed to be processor independent, so it shouldn’t matter whose chip is inside. But history suggests that it always matters.

Déjà vu All Over Again

I can’t help comparing the current net-book hoopla to the “network computer” buzz from a decade ago. Back then, everyone who wasn’t working for Intel predicted that the Internet would enable a whole new category of products that didn’t rely on x86 processors or Windows operating systems. We’d all read our e-mail, surf the Web, and run our applications from the big Internet cloud, blissfully unaware of whose processor or whose software was inside the box. The sun would shine, the angels would sing, and the Intel/Microsoft hegemony would be broken at last. More than a few serious business plans were based on this fairy tale, and more than a few chip companies went down to inglorious defeat because of it.

In the end, Intel and Microsoft maintained their dominance (indeed, they may have increased it), and network computers never came to be. The epic failure of NCs came down to two factors. First – ironically enough – was because customers didn’t care whose processor was in their machine. Unlike engineers, real customers had no particular axe to grind with Intel. They simply didn’t care about the religious wars going on among processor vendors and didn’t see the point. Whether the chip inside their computer came from Intel or Mattel, or was powered by a hamster in a treadmill, customers were supremely indifferent. So the strident “anti-x86” battle cry held no import for them whatsoever. It’s hard to lead a revolution when the populace is uninterested in politics.

Second, customers weren’t satisfied with the watered-down, generic applications available on NCs. Unlike processor chips, this was a difference they could see. Without the familiar plug-ins for Flash, Acrobat, Shockwave, and all the other features they’d grown accustomed to using on their PCs and Macs, NCs felt like weak substitutes. Although you could surf the Web or check your e-mail on a generic NC, nobody wanted to.

So here we are in 2009, and the arguments for net-book computers sound strangely familiar. The haves (Intel and Microsoft, mostly) find themselves under attack from the have-nots (every other microprocessor and operating system vendor). And again the population seems utterly indifferent to the pleas of the combatants. As laptop PCs get cheaper and cell phones get more capable, the supposed gap between those two products closes still further. What high-school or college student wants yet another device to tote around? An iPhone or BlackBerry costs around $200 (ignoring the monthly fees), while brand-new Windows XP laptops can be had for $500 or less… where’s the market window? A $200 net-book that doesn’t make phone calls and doesn’t fit in a pocket, but that still requires either a monthly data plan or round-the-clock proximity to someone else’s Wi-Fi network doesn’t make a lot of sense to me.

Whether the net-book market takes off or not, we can’t blame Freescale either way. The i.MX51 is a wonderful processor with high-end features and a low-ball price. At about $15, it’s a good choice for embedded developers working in any number of different product areas. It’s got a broad mix of I/O. It’s got a very fast processor core. It’s got good power management and can be throttled down to more moderate speeds if you don’t need 1-GHz performance. And maybe, just maybe, it will become a high-volume part that caught the wave of a new product category.

Leave a Reply

featured blogs
Nov 30, 2023
No one wants to waste unnecessary time in the model creation phase when using a modeling software. Rather than expect users to spend time trawling for published data and tediously model equipment items one by one from scratch, modeling software tends to include pre-configured...
Nov 27, 2023
See how we're harnessing generative AI throughout our suite of EDA tools with Synopsys.AI Copilot, the world's first GenAI capability for chip design.The post Meet Synopsys.ai Copilot, Industry's First GenAI Capability for Chip Design appeared first on Chip Design....
Nov 6, 2023
Suffice it to say that everyone and everything in these images was shot in-camera underwater, and that the results truly are haunting....

featured video

Dramatically Improve PPA and Productivity with Generative AI

Sponsored by Cadence Design Systems

Discover how you can quickly optimize flows for many blocks concurrently and use that knowledge for your next design. The Cadence Cerebrus Intelligent Chip Explorer is a revolutionary, AI-driven, automated approach to chip design flow optimization. Block engineers specify the design goals, and generative AI features within Cadence Cerebrus Explorer will intelligently optimize the design to meet the power, performance, and area (PPA) goals in a completely automated way.

Click here for more information

featured paper

Power and Performance Analysis of FIR Filters and FFTs on Intel Agilex® 7 FPGAs

Sponsored by Intel

Learn about the Future of Intel Programmable Solutions Group at intel.com/leap. The power and performance efficiency of digital signal processing (DSP) workloads play a significant role in the evolution of modern-day technology. Compare benchmarks of finite impulse response (FIR) filters and fast Fourier transform (FFT) designs on Intel Agilex® 7 FPGAs to publicly available results from AMD’s Versal* FPGAs and artificial intelligence engines.

Read more

featured chalk talk

VITA RF Product Portfolio: Enabling An OpenVPX World
Interoperability is a very valuable aspect of military and aerospace electronic designs and is a cornerstone to VITA, OpenVPX and SOSA. In this episode of Chalk Talk, Amelia Dalton and Eddie Alexander from Amphenol SV explore Amphenol SV’s portfolio of VITA RF solutions. They also examine the role that SOSA plays in the development of military and aerospace systems and how you can utilize Amphenol SV’s VITA RF solutions in your next design.
Oct 25, 2023