A (brief) primer on debugging data races — both parallel and distributed. Examples of common data races, and "data race smells"; coding techniques to avoid writing data races; testing techniques that work well for finding data races (and some discussion on why the usual testing does NOT do so well); applying statistical techniques to (psuedo-) random failures.
Program committee comment
Cliff is known for his work on efficient JIT compilation for JVM and many efficient concurrent and distributed algorithms, including the ones for machine learning. Being a distinguished expert in writing concurrent code, he will share the experience on how to detect and avoid data races in practice. Do not miss the chance to participate in this session and ask your questions personally!
Download presentation