Abstract: Local data-race-freedom guarantees, ensuring strong semantics for locations accessed by non-racy instructions, provide a fruitful methodology for modular reasoning in relaxed memory concurrency. We observe that standard compiler optimizations are in inherent conflict with such guarantees in general fully-relaxed memory models. Nevertheless, for a certain strengthening of the promising model by Lee et al. that only excludes relaxed RMW-store reorderings, we establish multiple useful local data-racefreedom guarantees that enhance the programmability aspect of the model.We also demonstrate that the performance price of forbidding these reorderings is insignificant. To the best of our knowledge, these results are the first to identify a model that includes the standard concurrency constructs, supports the efficient mapping of relaxed reads and writes to plain hardware loads and stores, and yet validates several local data-race-freedom guarantees. To gain confidence, our results are fully mechanized in Coq.

Talk and Poster Session Times

Click the button to the right of each event to participate at the scheduled time.