http://www.oopsla.org/2006/2006/index.php?option=com_content&task=view&id=66&Itemid=200

program registration submissions committee lodging portland

Method-Specific Dynamic Compilation using Logistic Regression

Method-Specific Dynamic Compilation using Logistic Regression

Research Paper

Wednesday, Oct 25, from 13:30 to 15:00

Determining the best set of optimizations to apply to a program has been a long standing problem for compiler writers. To reduce the complexity of this task, existing approaches typically apply the same set of optimizations passes to all procedures within a program, without regard to their particular structure. This paper develops a new method-specific approach that automatically selects the best optimizations on a per method basis within a dynamic compiler. Our approach uses the machine learning technique of logistic regression to automatically derive a predictive model which determines what optimizations to apply based on the features of a method. This technique is implemented in the Jikes RVM Java JIT compiler. Using this approach we reduce the average total execution time of the SPECjvm98 benchmarks by 29%. When the same heuristic is applied to the DaCapo+ benchmark suite, we obtain an average 33% reduction over the default level O2 setting.

John Cavazos, University of Edinburgh
Michael O'Boyle, University of Edinburgh

 
Research Papers in this Session

While Space Available
Search
program registration submissions committee lodging portland
For comments and questions about the web site
please contact us at support@oopsla.org
© 2005 OOPSLA