The electronics industry needs to come up with generic, easy-to-target platform processors if it is to capitalise on a current wave of innovation in processor design, according to the architect of the Transputer.
Professor David May also said that should such a platform emerge and achieve popularity, it could result in a new type of “fabless, hardware IP-less” firm that just works to tailor it to particular applications.
In a keynote address at a seminar on processor design in Cambridge this week, May said that approaches to achieving high performance computing at lower power need to be readily usable.
Current techniques such as automatic co-processor generation, parallel processing and large numbers of cores per chip, are powerful, but the difficulty of reliably generating and verifying complex heterogeneous designs could hamper their widespread uptake.
“There is a need for a generic kind of platform, whether they are FPGAs and chips full of processors, or whatever,” said May. He used the PC102 chip from Bath-based wireless processing firm PicoChip, which is a parallel array of heterogeneous elements, as an example.
“That’s not a general purpose array of processors, but it’s fairly flexible and you can imagine things that look like that which do have a fairly general purpose core,” said May. “I think alongside that you may well find that if the world shifts that way, and you get more of these generic platform technologies, then you’ll start to see an opportunity for companies that really do very little other than customise them. The sort of fabless, hardware IP-less company.”
May is currently head of the Department of Computer Science, University of Bristol.
www.cs.bris.ac.uk