High-Level Data-Access Analysis for Characterisation of (Sub)task-Level Parallelism in Java

In the era of future embedded systems the designer is confronted with multi-processor systems both for perfor-mance and energy reasons. Exploiting (sub)task-level par-allelism is becoming crucial because the instruction-level parallelism alone is insufficient. The challenge is to build compiler tools that support the exploration of the task-level parallelism in the programs. To achieve this goal, we have designed an analysis framework to evaluate the potential parallelism from sequential object-oriented programs. Parallel-performance and data-access analysis are the crucial techniques for estimation of the trans-formation effects. We have implemented support for platform-independent data-access analysis and profil-ing of Java programs, which is an extension to our earlier parallel-performance analysis framework. The toolkit com-prises automated design-time analysis for performance and data-access characterisation, program instrumenta-tion, program-profiling support and post-processing anal-ysis. We demonstrate the usability of our approach on a number of realistic Java applications. 1