The UCSB RACE Lab

Research on Adaptive Compilation Environments


People

Previous Students

Photo Album

---

Papers

Links to related
conferences

--- Location: Tr 935-101C
Find Us (Map)
---

Administration

 ---

Overview

The goal of our work is to improve the overall performance of programs using feedback from the executing program and underlying resource behavior. Our techniques adapt program and runtime components and executing code dynamically. We investigate various ways of collecting program and resource performance efficiently (profiling) as well as dynamic and adaptive compiler and runtime optimizations and specializations that exploit profile information. We consider a wide range of architectures from high-end cluster machines to mobile, resource-constrained, battery-powered devices. Recently, we have focused are efforts on emerging cloud computing systems. We investigate adaptive techniques for execution environments for Java, Microsoft CIL (the .Net Common Intermediate Language), and high-level dynamic languages, as well as for the Linux operating system and C & C++ runtimes. Below, we provide a list of links to web pages that provide additional information on the projects that we are currently working on. Please feel free to contact us for more information.

Projects

Software