John Mellor-Crummey's research focuses on software technology for high performance parallel computing. His ongoing research includes work on tools for measurement and analysis of applicatin performance, compiler and run-time technology for parallel and scientific computing, application performance modeling, and compiler technology for domain-specific languages. Past work has included developing techniques for execution replay of parallel programs, efficient software synchronization algorithms for shared-memory multiprocessors, and a system for efficiently detecting data races in executions of shared-memory programs using a combination of compile-time and run-time support.

In 2006, John Mellor-Crummey and Michael L. Scott were awarded the Dijkstra Prize in Distributed Computing for their paper Algorithms for Scalable Synchronization on Shared-Memory Multiprocessors, ACM Transactions on Computer Systems, February, 1991.