Written by HPC experts, this book provides a solid introduction to current mainstream computer architecture, dominant parallel programming models, and useful optimization strategies for scientific HPC. It facilitates an intuitive understanding of performance limitations without relying on heavy computer science knowledge. The authors highlight the importance of performance modeling of applications on all levels of a system's architecture and include numerous case studies drawn from their experiences in HPC user support, performance optimization, and benchmarking. Downloadable code and an annotated bibliography are available on the book's website.
