TI has just decided to show everyone what it can do with 28nm technology by pre-introducing its OMAP 5 mobile platform, the fifth generation of the successful OMAP series. So what can you do with 28nm? A lot, as it turns out. At its very “core,” the OMAP 5 platform features two ARM Cortex-A15 processors each capable of running at 2 GHz. However, two such engines clearly aren’t enough because there are also two ARM Cortex-M4 processor cores on the chip as well. The two ARM Cortex-A15 cores within the OMAP 5 platform are clearly application processors intended for some pretty advanced applications (keep reading), but what are you supposed to do with the two ARM Cortex-M4 processors? According to TI, they’re available to provide real-time response and to offload the ARM Cortex-A15 processors, with the net result that TI can claim a 60% power reduction “compared to a sample user experience on the OMAP 4 platform.”
What’s contributing to that reduction? Three things, it appears. First, the two ARM Cortex-A15 processors deliver 50% more computation per clock cycle compared to the two ARM Cortex-A9 processor cores used in the OMAP 4 platform. Next, the two ARM Cortex-M4 processor cores are available to offload real-time and periodic tasks from the larger, more power-hungry ARM Cortex-A15 cores, which makes it easier to reduce power consumption by moving certain frequently executed tasks to processing engines that consume less power per clock and by allowing the ARM Cortex-A15 application processors to operate at lower clock rates under less-than-peak load conditions.
Finally, the OMAP 5 platform incorporates several IP Subsystems (see the several recent EDA360 Insider blog posts on this topic). The included IP Subsystems further offload the OMAP 5 application processors, with dedicated engines for video, imaging and vision, DSP, 3D graphics, 2D graphics, display and security.
TI OMAP5432 Block diagram
Together, all of this computational power permits TI to project some fairly advanced apps for a future mobile communications device:
Next-generation natural user interfaces: gesture recognition for device control and for gaming.
Computational photography: to improve still and video imaging by compensating for the optical inadequacies of the small sensors and inexpensive lensing systems frequently incorporated into mobile handsets by using advanced image-processing apps such as camera stabilization, motion blur reduction, noise reduction, high dynamic range and face-based processing.
TI expects to sample OMAP 5 devices in the second half of 2011.