Intel (News - Alert) is moving more into the Internet of Things with its latest chip offerings, according to The Motley Fool. (Disclosure: I’ve contributed to Fool’s blogging network.)
The chipmaker has just launched the latest version of its Atom processor, the Apollo-Lake-I. The Atom series focuses on lower-power applications, which is perfect for embedded chips. Intel is targeting this version toward applications like medical devices and in-car entertainment systems.
The new system on a chip is intended to offer faster graphics performance with the Generation 9 graphics architecture. The Apollo-Lake-I offers an impressive 18 cores of graphics processors.
While better graphics are an obvious boon for graphical applications, such as in-car displays, they may not seem to offer a lot for other embedded IoT platforms that mainly rely on processing sensor data. In recent years, developers and chip makers have discovered that GPUs can offer surprising gains in performance for other kinds of processing, simply because they’re optimized for fast throughput.
System builders have built “desktop supercomputers” with multiple GPUs for scientific computing, allowing researchers to crunch experimental data without having to wait for time at supercomputing centers.
Apollo-Lake-I could offer the same kind of performance in a single chip, which could open up all kinds of applications.
With the desktop market saturated and many people opting for mobile devices powered by the ARM (News - Alert) processor, not x86, Intel is looking for new markets for its chips, and IoT looks like a good match. Embedded applications still aren’t so much a new application as much as Intel going full circle. The company’s earliest chips, including the 4004 and 8008 processors, were intended for things such as controlling traffic lights.
With modern processor technology, it’s possible to build systems that could monitor traffic flow across a city and intelligently control all the traffic lights to ensure smooth commutes, using the power of embedded GPU processing.
As chips continue to get cheaper and more powerful, IoT applications will get a lot smarter in the future.
Edited by Rory J. Thompson