Section: Software and Platforms

NLR

Participants : Alain Ketterlin, Philippe Clauss.

We have developed a program implementing our loop-nest recognition algorithm, detailed in [7] . This standalone, filter-like application takes as input a raw trace and builds a sequence of loop nests that, when executed, reproduce the trace. It is also able to predict forthcoming values at an arbitrary distance in the future. Its simple, text-based input format makes it applicable to all kinds of data. These data can take the form of simple numeric values, or have more elaborate structure, and can include symbols. The program is written in standard ANSI C. The code can also be used as a library.

We have used this code to evaluate the compression potential of loop nest recognition on memory address traces, with very good results. We have also shown that the predictive power of our model is competitive with other models on average.

The software is available upon request to anybody interested in trying to apply loop nest recognition. It has been distributed to a dozen of colleagues around the world. In particular, it has been used by Andres Charif-Rubial for his PhD work (Université de Versailles Saint-Quentin en Yvelines), and is now included in a released tool called MAQAO (http://www.maqao.org ). Our code is also used by Jean-Thomas ACQUAVIVA, at Commissariat à l'Énergie Atomique, for work on compressing instruction traces. These colleagues have slightly modified the code we gave them. We plan to release a stable version incorporating most of their changes in the near future. We also plan to change the license to avoid such drifts in the future.