The Third International Workshop on OpenMP, IWOMP 2007, was held at Beijing, China.This year'sworkshopcontinuedits traditionofbeingthe premier opportunity to learn more about OpenMP, to obtain practical experience and to interact with OpenMP users and developers. The workshop also served as a forum for presenting insights gained by practical experience, as well as research ideas and results related to OpenMP. A total of 28 submissions were received in response to a call for papers. Each submissionwasevaluatedbythreereviewersandadditionalreviewswerereceived for some papers. Based on the feedback received, 22 papers were accepted for inclusion in the proceedings. Of the 22 papers, 14 were accepted as full papers. We also accepted eight short papers, for each of which there was an opportunity to givea shortpresentationat the workshop, followed byposter demonstrations. Each paper was judged according to its originality, innovation, readability, and relevance to the expected audience. Due to the limited scope and time of the workshop and the high number of submissions received, only 50% of the total submissions were able to be included in the ?nal program. In addition to the contributed papers, the IWOMP 2007 program featured several keynote and banquet speakers: Trevor Mudge, Randy Brown, and Shah, Sanjiv. These speakers were selected due to their signi?cant contributions and reputation in the ?eld. A tutorial session and labs were also associated with IWOMP 2007.

This book presents a hybrid static-dynamic approach for efficient performance analysis of parallel applications on HPC systems. Performance analysis is essential to finding performance bottlenecks and understanding the performance behaviors of parallel applications on HPC systems. However, current performance analysis techniques usually incur significant overhead. Our book introduces a series of approaches for lightweight performance analysis. We combine static and dynamic analysis to reduce the overhead of performance analysis. Based on this hybrid static-dynamic approach, we then propose several innovative techniques for various performance analysis scenarios, including communication analysis, memory analysis, noise analysis, computation analysis, and scalability analysis. Through these specific performance analysis techniques, we convey to readers the idea of using static analysis to support dynamic analysis. To gain the most from the book, readers should have a basic grasp of parallel computing, computer architecture, and compilation techniques.