Scientists at the University of Massachusetts Lowell laugh in the face of Intel’s weedy handful of cores in its new CPU lineup: They’ve just squeezed over a thousand processor cores onto a single chip.
We’ve heard a lot about the potential for future desktop-sized supercomputers, but more than anything else this research proves that in the not-too-distant future it’s likely to be a reality. Interestingly enough, there’s also a green angle to this idea: FPGA chips can be more power efficient than their competitors, and if less computer time is needed to process complex tasks, then the overall power consumption of computers using the tech could be impressively low.
The advance was made by Dr. Wim Vanderbauwhede’s team, who programmed an advanced chip called a Field Programmable Gate Array (FPGA). FPGA systems have been around for a while, and their strength is that they can be programmed “in the field” to best suit whatever task they’re needed for, unlike the hard-coded silicon ship designs you’re probably imagining. The UM team’s innovation was in working out how to program the FPGA to act as mini processor cores, since the tech is typically difficult to work with. This has traditionally been a barrier to their use in desktop PCs, although small FPGAs are often found inside devices like LCD TVs.
Once the 1,000 individual CPU cores had been programmed onto the chip, the scientists took the necessary next step to prove how useful their innovation is: They ran an intensive algorithm through it to test how powerful it was, and they chose a tricky one too–at the core of motion MPEG video processing, used in many online video systems. The results speak for themselves. Using the kilo-core FPGA computer, the team was able to process 5 gigabytes per sec of movie files, which is about 20 times the rate that existing high-end computers can manage.