Since the start of the 2019-2020 academic year, Apple has been sponsoring VLSI and related Computer Hardware courses at UC Berkeley. In Spring 2021 I was enrolled in CS152, Computer Architecture and Engineering, which was one of these classes. In the course, we studied different techniques for designing computer hardware, and our labs consisted of running benchmarks on simulated RISCV hardware, and analyzing the results.
For Lab 3 (which involved executing a Spectre attack on an out of order RISCV processor) we were told that future labs would double as "Apple Design Competitions." This meant that the teams with the best benchmarks would be winning prizes sponsored by Apple. My team put a lot of extra time into these optimizations, often times reprogramming the same benchmarks multiple times, and heavily testing the smallest of changes to ensure we were always approaching optimal performance. We won two out of three competitions.
We Won: The fastest implementation of a perfect Spectre attack on the BOOM core (Lab 3) The fastest implementation of vectorized sparse matrix-vector multiply (Lab 4)
It was great to see our work acknowledged by the course staff and Apple. Every memeber of our group would meet multiple times a week, come up with ideas at all times of the day, and usually arrive at our brainstorm meetings with a laundry list of modifications we had already attempted. The best example of our teamwork is in Lab 4, where every member of our group had ideas that made it into the final code. I had so much fun working with my group in CS152. It was awesome to work with people who displayed so much enthusiasm for Computer Architecture, and I really hope I can work with them again
Also, thank you to Apple, for sponsoring this class and making this competition possible. I learned a lot through the competition--and the class being offered in the first place-- that I may not have otherwise.