Advanced Compiler and Runtime Support for Data Intensive Applications

• Renato Ferreira University of Maryland
• Gagan Agrawal University of Delaware
• Joel Saltz University of Maryland

DOI: https://doi.org/10.5753/sbac-pad.2000.41203

Resumo

Processing and analyzing large volumes of data play an increasingly important role in many domains of scientific research. However, high-level language and compiler support for developing such applications have been so far lacking. We are developing compiler and runtime techniques necessary to generate efficient implementations of such applications from a high-level programming language. Through advanced compiler analysis, our system is able to extract necessary information from the source. This information is used to generate an executable that makes extensive use of a runtime system called Active Data Repository (ADR). In this paper, we present advanced techniques thar can enhance the performance of compiler generated data intensive codes. We show how characterization of access patterns into dense and sparse and choosing appropriate execution strategy for each leads to better performance. We also show how demand-driven output space allocation on each processor, managed through a sparse data-structure, can reduce the number of tiles that need to be allocated on each processor and enhances performance.

Palavras-chave: data parallel compilers, data intensive applications, tiling, cluster of workstations

Referências

Anurag Acharya, Mustafa Uysal, Robert Bennett, Assaf Mendelson, Michael Beynon, Jeffrey K. Hollingsworth, Joel Saltz, and Alan Sussman. Tuning the performance of I/O intensive parallel applications. In Proceedings of the Fourth Annual Workshop on I/O in Parallel and Distributed Systems (IOPADS). ACM Press, May 1996.

Asmara Afework, Michael D. Beynon, Fabian Bustamante, Angelo Damarzo, Renato Ferreira, Robert Miller, Mark Silberman, Joel Saltz, Alan Sussman, and Hubert Tsang. Digital dynamic telephology - the Virtual Microscope. In Proceedings of the 1998 AMIA Annual Fall Symposium. American Medical Informatics Association, November 1998.

Gagan Agrawal, Renato Ferreira, and Joel Saltz. Language extensions and compilation techniques for data intensive computations. In Proceedings of Workshop on Compilers for Parallel Computing, January 2000.

Gagan Agrawal, Renato Ferreira, Joel Saltz, and Ruoming J in. High-level programming methodologies for data intensive computing. In proceedings of the Fifth Workshop on Languages, Compilers, and Run-time Systems for Scalable Computers, May 2000.

Francis Bodin, Peter Beckman, Dennis Gannon, Somasundaram Narayana, and Shelby X. Yang. Distributed pC++: Basic ideas for an object parallel language. Scientific Programming, 2(3), Fall 1993.

R. Bordawekar, A. Choudhary, K. Kennedy, C. Koelbel and M. Paleczny. A model and compilation strategy for out-of-core data parallel programs. In Proceedings of the Fifth ACM