COLLOQUIUM
             Computer Science Department, Boston University

Speaker: Professor Michael D. Smith
         Harvard University

Date:  Wednesday, February 8
Time:  3:00
Place: Room MCS 135, 111 Cummington Street 
       (for directions, see www.cs.bu.edu/colloquium)

Title: Improving Region Selection in Dynamic Optimization Systems

Abstract:

The performance of a dynamic optimization system depends heavily on the
code it selects to optimize. Many current systems follow the design of HP
Dynamo and select a single interprocedural path, or trace, as the unit of code
optimization and code caching. Though this approach to region selection
has worked well in practice, we show that it is possible to adapt this basic
approach to produce regions with greater locality, less needless code
duplication, and fewer profiling counters. In particular, we propose two
new region-selection algorithms and evaluate them against Dynamo's selection
mechanism, Next-Executing Tail (NET). Our first algorithm, Last-Executed
Iteration (LEI), identifies cyclic paths of execution better than NET,
thus improving locality of execution while reducing the size of the code
cache. Our second algorithm allows overlapping traces of similar execution
frequency to be combined into a single large region. This second
technique can be applied to both NET and LEI, and we find that it significantly
improves metrics of locality and memory overhead for each.

This work was done in cooperation with David Hiniker and Kim Hazelwood.

Biography:

Michael D. Smith (smith@eecs.harvard.edu) is a Gordon McKay Professor of
Computer Science and Electrical Engineering and Associate Dean for
Computer Science and Engineering in the Division of Engineering and Applied
Sciences. His research interests include dynamic optimization, machine-specific
and profile-driven compilation, high-performance computer architecture, and
practical applications of security.  Mike is also Chief Scientist and
co-founder of Liquid Machines, Inc., a provider of unique solutions for
enterprise rights management.  He is a member of the IEEE and ACM
professional societies, and holds degrees from Stanford University, WPI,
and Princeton University.

Host: Azer Bestavros