The year 2012 saw continued growth in the adoption of mobile communications gadgets, but few technological breakthroughs. But in the months ahead, the guts of mobile devices could be ripe for technological disruption and significant performance boosts.
Smartphones and tablets are, in some respects, primitive and inefficient inside. Crack open their sleek surfaces and you find a jumble of standalone processors. There are typically several modems for processing for different wireless standards: 2G, 3G, 4G LTE, Wi-Fi, Bluetooth, GPS, and, in newer phones, near-field communications. Finally, there are processors for image and video processing, and for the accelerometer and gyrometer.
Housing and powering all of these standalone islands of processing is fundamentally inefficient, but technologies are emerging for putting them all together. Coherent Logix, a startup based in Austin, Texas, is developing a single processor that, the company says, can handle all of the above with functions nimbly controlled by software that can switch between functions in a few milliseconds.
Chips that process software versions of radio components that were once only implemented as hardware—such as filters, amplifiers, and modulators—have been around for years. So has the ability for such chips to be reprogrammable to handle different frequencies. But adding everything else—such as image and video processing—to the same processor and making it work in a smartphone would be quite new. “We believe it’s not enough just to put the radios together and put them in software. To get the most powerful, flexible, dynamic system, you want to tie all the video and photo processing together with the radio processing,” says Michael Doerr, the company’s founder and chief technology officer. “And we have a unique processor that can do that, but it’s a question of getting to the right performance and power trade-off.”
The company has supplied military customers with chips necessary to make sophisticated programmable radios, but the technology has been held back from the mass market by its cost, power requirements, and programming complexity. Now the company says it is working with one of the top five smartphone makers—it won’t name which—on a handset version to be prototyped by 2014 using a next-gen version of its programmable processor. If it succeeds, he says, “It will be a giant leap for the industry. Nobody has come out with a device like this.”
The company has some tentative validation in the research community. “It will be interesting to see how far they will be able to lower the power consumption in the near future, but the potential exists for supercomputing performance on a handset,” says Jeff Reed, who directs the wireless research institute at Virginia Tech and is familiar with the technology. Success would be “game-changing” because the resulting device—if not literally as powerful as a supercomputer—would certainly have the capabilities of today’s powerful laptop or desktop machines without consuming more power.
What’s more, devices might be useful for longer, because the hardware would not become obsolete nearly as quickly, and more upgrades could be done with software to, for example, add a new kind of image or video processor. “What makes them well respected is their software tools, which make multiprocessor processing much easier than in the past,” Reed adds.
Coherent Logix isn’t the only player in the game, of course. Major chip makers have steadily been improving performance and driving down power consumption, while entrants such as Tilera and Mindspeed are also working on better programmable processors.
In the meantime, Coherent Logix’s chips are headed to base stations, which stand to gain from similar boosts in programmability and flexibility. Glance at the side of an urban building and you’ll likely see several transmitter boxes. Each handles a different frequency for each of three or four or more carriers.
The Coherent Logix programmable chip is being adopted by Public Wireless, a vendor that provides base stations for the wireless industry. Changes to wireless protocols require physical visits to base stations, which increasingly consist of larger numbers of small transmitters called small cells. As more and more of these are installed, “the sheer number of locations and costs of getting to those locations over time is fairly significant,” says Rob Reagan, founder of Public Wireless.
He argues that the cheapest way to deal with this is to send new software to the base station to redefine what the radio processor should be doing. In theory, the wireless companies could even share the same box, much as the same power line is shared by different electricity providers. “The whole space is ripe for significant disruption,” Reagan adds.
Wireless carriers are also developing technologies including small cells that can more efficiently handle the expected expansion of LTE networks (see “Tiny Transmitter Could Help Avert Data Throttling”). AT&T last year said it would help expand capacity in part by adopting small cells (see “The Spectrum Crunch That Wasn’t”), joining Verizon, Sprint, Telefónica, China Mobile, Vodafone, and many others. Major base station vendors such as Ericsson, Alcatel-Lucent, Nokia, and Siemens are also evolving their products.