J. Nelson Amaral
University of Alberta, Canada
Distinguished ACM Speaker

14 de julho – 10:00h - Auditório do CIC/UnB (Módulo 18 – Subsolo ICC)

ABSTRACT: The compilation of complex programs for efficient execution in modern computer architectures requires that many difficult optimization problems be addressed. Recently there has been great interest in the application of automated learning techniques to improve the compilation process. In this talk I report on our recent
experience investigating the use of Support Vector Machines (SVMs) to improve compilation strategies in Testarossa, the commercial Just-in-Time compiler that is used in IBM's J9 Java Virtual Machine.

This talk not only presents the results of applying SVMs to this learning task, but also discusses very important practical issues, such as the precise measurement of compilation and execution time and the necessary reduction in the volume of data used for training, that arise when integrating a machine-learned model into a commercial
compilation environment. The machine-learned plans outperform the original Testarossa for start-up performance, but not for throughput performance, for which Testarossa has been highly hand-tuned for many years.

ABOUT THE SPEAKER: José Nelson Amaral is a professor of Computing Science at the University of Alberta, Canada. He received the Ph.D.in Electrical and Computer Engineering from the University of Texas at Austin, in 1994, the M.E. from the Instituto Tecnológico de Aeronautica, São José dos Campos, SP, Brazil, and the B.E. from the
Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS), RS, Brazil.  His current research interests include Compiler Design, Static Analysis, Feedback Directed Compilation, Computer Architecture, High-Performance Computer Systems, and the application of learning methods to the design of compilers.  His previous research includes Cache-Conscious Algorithms, Internet Protocol Routing Caches, Artificial Neural Networks, Combinatorial Optimization Problems,
Parallel Architectures for Symbolic Processing, Multi-Threaded Architectures and Programming Models. Dr. Amaral is a Senior Member of the IEEE and a Distinguished Speaker for ACM

Diretoria
IEEE Seção Brasília