UTCS Colloquia, Shan Lu, "An Effect-Oriented Approach to Concurrency-Bug Detection and Recovery"

Contact Name: 
Katie Dahm
GDC Auditorium
Mar 18, 2014 11:00am - 12:00pm
Lorenzo Alvisi
Speaker Affiliation: 
University of Wisconsin, Madison

Signup Schedule: http://apps.cs.utexas.edu/talkschedules/cgi/list_events.cgi

Talk Audience: UTCS Faculty, Grads, Undergrads, Other Interested Parties

Host:  Lorenzo Alvisi

Talk Abstract: Concurrency bugs are notoriously difficult to detect due to their complicated root causes --untimely interactions among threads. With the pervasive adoption of multi-core machines and multi-threaded software, undetected concurrency bugs are widespread in the field and lead to severe failures on user machines.

This talk will present an effect-oriented approach to fighting concurrency bugs. I will first use a characteristics study to demonstrate that concurrency bugs have simple effect patterns, very similar to those of sequential bugs, in spite of their complicated root causes. I will then present ConMem and ConSeq, two dynamic bug-detection tools that leverage these effect patterns to achieve high coverage and high accuracy. Next, I will discuss ConAir, a static code-transformation tool that leverages these effect patterns to enable automated failure recovery with negligible run-time overhead on commodity machines.

I will conclude the talk by discussing other research in my group that tackles concurrency bugs and performance bugs.

Speaker Bio: Shan Lu is the Clare Boothe Luce Assistant Professor of Computer Sciences at University of Wisconsin, Madison. She earned her Ph.D. at University of Illinois, Urbana-Champaign, in 2008. At University of Wisconsin, her group works on detecting, diagnosing, and fixing concurrency bugs and performance bugs. Shan Lu won Alfred P. Sloan Research Fellow in 2014, Distinguished Alumni Educator Award from Department of Computer Science at University of Illinois in 2013, and NSF Career Award in 2010. Her co-authored papers won the Best Paper Award at USENIX FAST in 2013, ACM-SIGPLAN CACM Research Highlight Nomination in 2011, and IEEE Micro Top Picks in 2006. She currently serves as the Information Director of ACM-SIGOPS.