Original press release was issued by UC Davis, writen by Andy Fell.
The recently unveiled microchip created by researchers at University of California, Davis is the world’s first chip that hosts 1000 independently programmable processors. The “KiloCore” contains 621 million transistors, and its reported maximum computation rate is at 1.78 trillion operations per second. Even if you aren’t a computer engineer, you can probably guess that those numbers are nothing short of impressive – but what do they actually mean?
Each processor core can run its own small program independently of the others, which is a fundamentally more flexible approach than so-called Single-Instruction-Multiple-Data approaches utilized by processors such as GPUs; the idea is to break an application up into many small pieces, each of which can run in parallel on different processors, enabling high throughput with lower energy use, said Bevan Baas, professor of electrical and computer engineering, who led the team that designed the chip architecture.
Yes, technically, chips with this number of processors are already available today – in the said GPUs – but with a significant difference. While GPUs direct the collective processing power from a central controller, KiloCore’s independent cores are designed to run different computer programs in parallel. Its architecture thus allows for immensely more complex computations.
“Perhaps thinking of the chip as a room filled with 1000 computers is helpful, though a stretch,” Bevan Baas, a computer engineering professor and leader of the UC Davis team that developed the chip, told Baas to Motherboard.
But raw power is not where the advantage of this kind of multiple-core architecture ends. Since the processors really are independent, they can shut themselves off when not needed, said graduate student Brent Bohnenstiehl, who developed the principal architecture. That makes KiloCore the most energy-efficient “many-core” processor ever reported, Baas said. For example, the 1,000 processors can execute 115 billion instructions per second while dissipating only 0.7 Watts, low enough to be powered by a single AA battery. The KiloCore chip executes instructions more than 100 times more efficiently than a modern laptop processor.
Cores operate at an average maximum clock frequency of 1.78 GHz, and they transfer data directly to each other rather than using a pooled memory area that can become a bottleneck for data. And that is the third major difference between KiloCore and your average computer – it does not use RAM.
Applications already developed for the chip include wireless coding/decoding, video processing, encryption, and others involving large amounts of parallel data such as scientific data applications and datacenter record processing.
The team has completed a compiler and automatic program mapping tools for use in programming the chip.