ACE: Adaptive Compilation Environments

The ACE Project proposes to fundamentally change the way in which compilers are developed, replacing hand-tuning by self-tuning compilers that adapt automatically to match the characteristics of each target computing system. By automating the process of performance tuning, programmers will have more time to focus on higher level issues, e.g., algorithm design, robustness, and correctness. Moreover, intelligent compilers will reduce the need for specialized expertise across the range of targetted computing systems.

 Selected Publications:

[1] Intelligent Compilers.
John Cavazos.
IWAPT 2008 [gzip'd PS]  [PDF]  [SLIDES]

[2] Rapidly Selecting Good Compiler Optimizations using Performance Counters.
John Cavazos, Grigori Fursin, Felix Agakov, Edwin Bonilla, Michael F. P. O'Boyle, and Olivier Temam. Acceptance: 26/84 (31%)
CGO 2007 [gzip'd PS]  [PDF]

[3] Fast Compiler Optimisation Evaluation Using Code-Feature Based Performance Prediction.
Christophe Dubach, John Cavazos, Bjorne Franke, Grigori Fursin, Michael F. P. O'Boyle, and Olivier Temam. Acceptance 28/56 (50%)
Computing Frontiers 2007 [gzip'd PS]  [PDF]

[4] Method-Specific Dynamic Compilation using Logistic Regression.
John Cavazos and Michael F. P. O'Boyle. Acceptance: 26/157 (16%)
OOPSLA 2006 [gzip'd PS]  [PDF]