Having Fun as an MIT Professor: multiprocessors and wireless chips
April 25, 2022
by Lance

During my time on the M.I.T. faculty, my students and I did get some chances to do some work that was just plain fun. In 1988 Charles Zukowski was taking an advanced math course and we decided to apply it to massively parallel multiprocessors. We did what electrical engineers had done for 100 years, we went to the continuous limit. The paper is Continuous Models for Communication Density Constraints on Multiprocessor Performance. That was theory.

deldot

 

On the experimental side, Adam Malamy, Charlie Selvidge, and I built a monolithic digital CMOS device that connected to the outside world without wires. Power, signal in, and signal out, were all accomplished wirelessly. It was actually pretty tricky to do with the technology of the day: 3 micron CMOS with Aluminum metalization. The paper was A magnetic power and communication interface for a CMOS integrated circuit.

 

magcomm