The research focuses on the design and implementation of a framework, called the ADaptive Object RE-optimization (ADORE) system, to evaluate the effectiveness of dynamic binary re-optimization, and to study the interaction between architecture/micro-architecture and a dynamic binary re-optimizer. The topics studied include: efficient profiling techniques; runtime hot region selection and formation; light-weight runtime optimizations; code patching and region linking techniques; architectural and micro-architectural support for dynamic binary optimization; and runtime specialization and value profiling. The first implementation of the ADORE system will be based on two different implementations of the IA-64 architecture, the Itanium and the McKinley. The performance of application code optimized for Itanium will be tested on McKinley using the ADORE system.

Agency
National Science Foundation (NSF)
Institute
Division of Computer and Communication Foundations (CCF)
Type
Standard Grant (Standard)
Application #
0105574
Program Officer
Pratibha Varma-Nelson
Project Start
Project End
Budget Start
2001-09-01
Budget End
2004-08-31
Support Year
Fiscal Year
2001
Total Cost
$306,173
Indirect Cost
Name
University of Minnesota Twin Cities
Department
Type
DUNS #
City
Minneapolis
State
MN
Country
United States
Zip Code
55455