Just in Time Load Balancing

摘要: Leveraging Loop Level Parallelism (LLP) is one of the most attractive techniques for improving program performance on emerging multi-cores. Ordinary programs contain a large amount parallel and DOALL loops, however multi-core designs feature rapid increase in number on-chip cores ways such share resources - as pipeline memory hierarchy, leads to an possible high-performance configurations. This trend design makes attaining peak through exploitation LLP increasingly complex problem.