Consumers who think six or 8 cores are many in a processor probably never had a reason to look up parallel chips and digital signal processors (DSP), but French company Kalray may change that.
What they've announced is a DSP chip which, thanks to its 256 individual cores (yes, that many) can perform 500 billion operations per second (200 GigaFLOPS).
One would think that something like that would require a lot of power, but that does not apply here. The processor actually makes do with 4W, less than the average mobile chip.
For comparison, a mighty Tesla
series GPU has a performance of 1,030 to 2,280 (single-precision). A factor of 5-10 isn't bad at all really.
Granted, being a digital signal processor, Kalray's chip (MPPA 256) has a narrower range of uses (accelerating multimedia codecs and filters) than, say, CPUs (can process pretty much anything) and GPUs (are best at parallel tasks).
To make things even more interesting, the newcomer is manufactured on the 28nm manufacturing process from Taiwan Semiconductor Manufacturing Company
(TSMC), the same foundry that makes those Tesla chips mentioned above, and all the other GPUs that NVIDIA uses.
Still, that doesn't change the fact that this is one strong chip that may endanger other many-core projects, like Adapteva's Parallella “supercomputer.”
"Kalray's technology has been developed with many OEM partners across several vertical markets, as well as through partnering with third party software vendors,” says Joël Monnier, CEO.
“Our first processor achieves a global processing power of 500 billion operations per second, along with a much lower power consumption than competitive solutions. Embedded designers will get the benefit of high processing power, low power consumption and high level programming to develop innovative applications in the fields of image processing, signal processing, control, communications and data security. The access cost of MPPA processors makes them optimum for all low to medium volume applications.”
Kalray expects to have customers from the fields of image processing, signal processing, control, communications and data security. Go here
to read more.