The deterministic user-level replay of concurrent programs

Speaker:        Dr. Charles ZHANG
                Department of Computer Science and Engineering
                Hong Kong University of Science and Technology

Title:          "The deterministic user-level replay of concurrent
                 programs"

Date:           Monday, 5 November 2012

Time:           4:00pm - 5:00pm

Venue:          Lecture Theatre F (near lifts 25/26), HKUST

Abstract:

A user-level deterministic log/replay system is able to faithfully
re-enact a past multithreaded execution without any modifications to the
underlying OS or hardware. As concurrent software gets increasingly
popular, such capability is highly useful and sometimes pivotal in
understanding and fixing concurrency bugs. In this talk, I will present
some of the major research challenges of the user-level replay together
with three techniques, developed by our research group, that make
significant improvements over the prior art. I will first describe LEAP,
an order-based replay tool, that works for many popular multithreaded Java
programs. I will then describe STRIDE, a hybrid technique that achieves a
lower recording overhead by permitting the concurrent-read-exclusive-write
(CREW) semantic. Finally, I will present CLAP, a search-based and light
weight replaying technique for C/C++ programs, that uses SMT solvers and
thread-local profiles to achieve the replay determinism.

*******************
Biography:

Charles Zhang is an assistant professor in the department of computer
science and engineering at the Hong Kong University of Science and
Technology. His major research area is software engineering with the
current focus on concurrent systems. He has published extensively at
premium conferences and journals such as TOSEM, TSE, IEEE TPDS, ICSE, FSE,
ISSTA, OOPSLA, ECOOP, ACM/USENIX MIDDLEWARE, and AOSD. He has served on
the organizational and technical committees of OOPSLA, FSE, ISSTA, PLDI,
AOSD, and APSEC. Charles obtained his Ph.D, M.Sc, and B.Sc. with honors,
all from University of Toronto.