Are you listening to these words as a podcast on the device which used to be your cell phone and now has become the victim of convergence combining a myriad of consumer functions in a single unit? Are you wondering what the next level of convergence will bring?
Be advised that the electronics industry is at risk of not being able to deliver that next generation. That's because methodologies to develop cheaper, faster, more power-efficient and reliable chips that drive consumer electronics are simply running out of steam. Innovative thinking is required to address design and verification across multiple specializations hardware, software, architecture and verification.
Examples of the first generations of platforms targeted to consumer markets are Philips' nExperia, Texas Instruments' OMAP and STMicroelectronics' Nomadik. On the nExperia silicon, users already find three processors and about 20 dedicated hardware components. Given the huge investment necessary for a platform easily ranging between $20 and $100 million, not only is time to market important but time in market has to be optimized to allow proper returns on the investment. This in turn has driven the trend to enable more flexibility using software on processors.
However, power consumption has since become the limiting factor. While the desktop computing world has stopped at about 4 GHz, processors in the embedded consumer world have reached 1 GHz only to run into similar power and clock frequency limitations. Using 10 processors instead, each of them consuming 100 times less power, leads to an overall 10 times reduction in power consumption. In reality, the industry is lacking methodologies to efficiently parallelize and distribute the software across those processors in an automatic fashion.
The current situation suggests great discontinuity in the electronics industry short term and, in turn, offers a tremendous long-term opportunity for those willing to seize it. As the industry moves forward, developing and programming multi-processor systems-on-chip (MPSoC) will require a unified systems design automation approach in which hardware and software technologies and design processes are combined in a seamless development environment.
This change will also cause a restructuring of the electronics industry as business units like the one developing nExperia have already become software companies with the number of software developers exceeding the number of hardware designers. And, this trend will become more prominent with the average number of processors per MPSoC growing and programmers finding that their existing tools do not work for the new class of multi-processor or multi-core chips.
Assessing who will be able to seize this opportunity requires understanding of the fundamental requirements a software development environment for MPSoCs has to meet. Software developers can no longer wait for the chip to debug the software because the chip is becoming the software. Hardware developers realize that the MPSoC, its structure and partitioning have become part of the software problem.
Not quite meeting the requirements of an application market has unfortunate consequences for a silicon platform given the development effort. Even more flexibility is required leading to more processors and cores. The number of processors and in which configuration and partitioning is dependent on the software requirement running on it.