Efficient Deterministic Replay and Data Race Detection for Multithreaded Programs

Abstract

With the advent of multiprocessor systems, it is now the role of the programmers to explicitly expose parallelism and take advantage of parallel computing resources. However, multithreaded programs are vulnerable to concurrency errors such as data races. To help programmers develop more reliable and secure concurrent programs, this talk presents DoublePlay, a deterministic replay system; and TxRace, a data race detector.

DoublePlay is an efficient OS-based deterministic replay system that records and reproduces the execution of multithreaded programs on multiprocessors. DoublePlay introduces a new insight that it is sufficient for many replay uses to guarantee only the same output and the final states between the recorded and replayed executions, and thus it is possible to support replay at a low cost.

TxRace is a light-weight dynamic data race detector that leverages commodity hardware transactional memory (HTM) to speed up data race detection. TxRace exploits the observation that the conflict detection mechanism of HTM can be repurposed for data race detection; and addresses challenges in using the current best-effort commodity HTMs.

Short bio

Dongyoon Lee is an Assistant Professor in the Computer Science department at Virginia Tech. He received the M.S. (2009) and Ph.D. (2013) degrees in Computer Science and Engineering at the University of Michigan, Ann Arbor; and the B.S. (2004) degree in Electrical Engineering from the Seoul National University in Korea. Before joining VT, he worked as an academic visitor in the next generation middleware platforms department at IBM T. J. Watson Research Center. He has been awarded Google Research Award; ProQuest Distinguished Dissertation Award; VMWare Graduate Fellowship; and the best paper at ASPLOS 2011.

Resources

• talk slide