Lock Contention Management in Multithreaded MPI

Abdelhalim Amer,Huiwei Lu,Pavan Balaji,Milind Chabbi,Yanjie Wei,Jeff Hammond,Satoshi Matsuoka

DOI: https://doi.org/10.1145/3275443

2019-01-23

ACM Transactions on Parallel Computing

Abstract:In this article, we investigate contention management in lock-based thread-safe MPI libraries. Specifically, we make two assumptions: (1) locks are the only form of synchronization when protecting communication paths; and (2) contention occurs, and thus serialization is unavoidable. Our work distinguishes between lock acquisitions with respect to work being performed inside a critical section; productive vs. unproductive . Waiting for message reception without doing anything else inside a critical section is an example of unproductive lock acquisition. We show that the high-throughput nature of modern scalable locking protocols translates into better communication progress for throughput-intensive MPI communication but negatively impacts latency-sensitive communication because of overzealous unproductive lock acquisition. To reduce unproductive lock acquisitions, we devised a method that promotes threads with productive work using a generic two-level priority locking protocol. Our results show that using a high-throughput protocol for productive work and a fair protocol for less productive code paths ensures the best tradeoff for fine-grained communication, whereas a fair protocol is sufficient for more coarse-grained communication. Although these efforts have been rewarding, scalability degradation remains significant. We discuss techniques that diverge from the pure locking model and offer the potential to further improve scalability.