Does this chip make my 1000 cores look fat?

A report from the University of Glasgow heralds a research breakthrough where over 1000 cores were built into a single chip. Field Programmable Gate Array (FPGA) technology was used. The chip was reported to perform some graphics related computation at a rate of 20X current processors running through data at a rate of 5Gb per second.

It's a proof-of-concept effort for now; Dr. Vanderbauwhede, who heads up the research effort, hopes to present this research at the International Symposium on Applied Reconfigurable Computing in March.

I could find no indication of how the cores operated or how they were programmed. I assume that these were fairly simple cores and that they may have used some form of data parallel of SIMD execution. Even so, just knowing that it would be possible to construct such chips sparks the imagination.

There have been large collections of very simple processors that have been used to solve important problems.  The Thinking Machines CM-1 and CM-2 come to my mind. I would also add our brains to the list of systems that are made of a large collection of simple processors. We may soon be able to execute brute force parallel computations that can mimic intelligence at speeds close to real-time.
For more complete information about compiler optimizations, see our Optimization Notice.