Continuous adaptive object-code re-optimization framework
Journal
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Journal Volume
3189
Pages
241-255
Date Issued
2004
Author(s)
Abstract
ynamic optimization presents opportunities for finding run-time bottlenecks and deploying optimizations in statically compiled programs. In this paper, we discuss our current implementation of our hardware sampling based dynamic optimization framework and applying our dynamic optimization system to various SPEC2000 benchmarks compiled with the ORC compiler at optimization level O2 and executed on an Itanium-2 machine. We use our optimization system to apply memory prefetching optimizations, improving the performance of multiple benchmark programs. © Springer-Verlag 2004.
Other Subjects
Benchmarking; Benchmark programs; Compiled programs; Dynamic optimization; Object code; Optimization framework; Optimization levels; Optimization system; Sampling-based; Embedded systems
Type
journal article