News‎ > ‎

Mobile Platform is can solve complex problems

posted 17 May 2010, 02:39 by Ramy Caspi   [ updated 17 May 2010, 02:42 ]
Today I saw a short video that shows how powerful the mobile handset continues to become, in this case solving a Rubik’s Cube in seconds by using image capture, advanced algorithms and a powerful mobile platform, a Motorola Droid. The Droid is based on the TI OMAP 3430 application processor which is based on the ARM Cortex A8 processor.

ARM "rubik"

(Update 05/12: Wired Gadget Lab reported that the video has been temporarily removed by ARM to make some changes and that it is expected to be back online by 05/14)

This video triggered some thoughts about the future of the mobile handset as a platform to solve complex computational problems…

Imagine a future where battery consumption and/or network coverage were not an issue as today, and where sets of mobile handsets can be federated in real-time; hundreds, thousands or millions of handsets, as they become available or not (on the network), all working “together” to solve specific complex computational problems, in a similar fashion to how SETI @ Home uses millions and millions of idle CPU cycles from thousands (millions?) of PCs from around the world.

Let’s look at a relevant scenario. While there are millions of Facebook users there are billions of mobile handsets/users; now imagine a time in the future where most of the handsets are powerful-enough to solve the relevant parts of “their” social graph; millions of individual handsets solving their owner’s social graphs in real-time as it changes, then all stitched together somehow on the edge (the handset) or centrally on the network.

“Availability” is what precludes this vision of mobile-based distributed computing from happening; either because the handsets are not powerful-enough today or because of lack of connectivity due to issues related to battery, network coverage or IP visibility; but these limiting factors will go away, in the future.

The mobile handset is becoming so powerful that what I described above might be closer to reality than what we think.

The OMAP 3430