Ask Me Anything with Kathryn McKinley
summary
Live Q&A explains why the DaCapo suite was born from an NSF review, the DaCapo 20-author process and strict criteria (no user interaction, commercial relevance, runnable in Harvest). Practical advice covers mentoring via regular internships and open-source collaborations, balancing leadership with research, and structuring industry–academic IP. Technical claims: datacenter workloads make GC economics in cloud essential because leaking memory raises operating cost, Rust reduces bugs at the cost of programmer burden, and AI lets researchers reframe problems (example: ML-driven huge‑page allocation).
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:34Can you talk a little bit about the DeCapo benchmark suite
- 4:07What advice would you give PhD graduates who are going to work in the industry
- 5:51How do you pick research topics
- 9:12Why do you do both
- 14:07Google walkout
- 17:15Garbage collection
- 21:12Leadership
- 23:59Vertical Integration
- 27:05Collaborations
- 31:57Outro
