Running a Hands-On Laboratory Course Remotely

FeaturedHardwareComputation
Written By Phillip Stanley-Marbell
Two iCE40 MDP Evaluation Boards setup for measurement of the current drawn on the VCC_ICE_D supply rail (FPGA core supply rail for the fourth FPGA on the MDP).

Two iCE40 MDP Evaluation Boards setup for measurement of the current drawn on the VCC_ICE_D supply rail (FPGA core supply rail for the fourth FPGA on the MDP).

To give participants the opportunity to still carry out the power measurement component of the RISC-V Processor Design Project when running the project completely online, I setup a few iCE40 MDP evaluation boards for power measurement in my workspace. Participants provide FPGA bitstreams during live sessions which we hold twice a week and they get to see their measurements happening on live video.

I’ve set things up so that I can overlay the output of an oscilloscope and logic analyzer (Tektronix MDO4104C) over the live video stream, so we can have a seamless combination of the discussion and slides and measurements.

Overlaying live oscilloscope output over the slides during the live “online laboratory” sessions in which participants submit their FPGA bitstreams and I load them to the FPGA and measure the FPGA’s power dissipation on their behalf. Each bitstream -> configuration -> measurement step only takes about a minute and a half.

Quality of the video stream is a challenge when participants (and myself) are on potentially-ropy network connections. I’ve tried pushing the video stream directly into a video call in previous sessions, but there’s a much better alternative. More on that shortly…

Phillip Stanley-Marbell

Currently: I am an Associate Professor in the Electrical Engineering Division of the Department of Engineering at the University of Cambridge, a Faculty Fellow at the Alan Turing Institute, and the founder of Signaloid, a startup developing a new approach to computation that interacts with the physical world. At Cambridge, I lead the Physical Computation Laboratory. At the Turing I co-lead the Interest Group on Resource- and Data-Constrained AI.

My research explores how to exploit the structure of signals in the physical world and the flexibility of human perception to make computation more efficient. Some of my ongoing work applies this idea to new hardware architectures for tracking uncertainty in computation and sensing and new methods for learning models from physical sensor data.

Previously: Ph.D., 2007, CMU. I spent 2006–2008 at Technische Universiteit Eindhoven in the Netherlands, joined IBM Research in Zürich, Switzerland, as a permanent Research Staff Member from 2008–2012, and then joined Apple in Cupertino from 2012–2014. I moved back to academia in 2014: I was in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) from 2014-2017 and joined the University of Cambridge as a faculty member in 2017.

http://phillipstanleymarbell.org
Previous

What Students Say About the Course