Long-form writing from members of the lab — research notes accompanying recent papers, engineering write-ups, and the occasional reflection on the practice of running an architecture group.
Silicon photonics promises terabit-per-second on-chip links. Process variation is what stands between that promise and the wafer. Here’s what we measured, and what we did about it.
Microring resonators want a clean thermal environment. The real world does not provide one. A closed-loop calibration scheme that fits inside the link, not on the bench.
Quality-configurable arithmetic units let an embedded neural network spend bits where they matter and save them where they do not. A short tour of what changes when accuracy is a knob.
Keyword-spotting accelerators are typically on, always polling. Spiking accelerators only do work when something interesting happens. A look at what that buys.
Five years after moving the group from École Centrale de Lyon to Concordia, a brief inventory of what travelled, what did not, and what we built locally.