Exploring concurrency using the parallel analysis tool

Abstract

Laboratory work in the CS classroom is intended to solidify essential concepts and core design principles. Because of recent advances and the widespread adoption of multicore technologies, one area of investigation that has become increasingly important across all levels of CS instruction is parallel computing. This paper describes the initial version of the Parallel Analysis Tool (PAT), a pedagogical tool designed to assist undergraduate students in visualizing concurrency and effectively connecting parallel processing to applied coding strategies. The PAT is a complete Java development environment, with an emphasis on (1) helping students to identify appropriate code locations where parallelization can be applied and (2) allowing students to subsequently examine the practical performance tradeoffs of these parallelization decisions in a laboratory setting. The Parallel Quotient (PQ), a fundamental dimensionless metric generated by the PAT, supports the student's conceptual understanding and analysis of the relative benefits of employing various parallel programming strategies.

E-learning and Benchmarking Platform for Parallel and Distributed Computing

“…Rague [11] introduced the Parallel Analysis Tool (PAT), a pedagogical tool, which assists the undergraduate students to transform a specific code section from sequential to parallel. The students can examine the impact of parallelizing a block of code at a time on overall program performance.…”

Parallel from the beginning

“…A parallel language and an interactive programming environment for teaching elementary school students are introduced in [5]. A parallelism analysis tool for Java is introduced in [16]. The primary intention is to allow students to find parallelism within Java source code.…”