Mastering adaptive mesh refinement for parallel processors shows how to run complex simulations efficiently. It explains the key ideas behind AMR and how to balance work across machines.
This book introduces AMR as a practical method for solving problems that need variable resolution. It covers how nested grids work, how refinement ties to time stepping, and how data and boundary communications are managed to keep calculations accurate and stable. The discussion includes real‑world implementation details on Cray XMP and Ultracomputer systems, along with strategies to improve performance.
- How adaptive grids are created and refined to target high‑error regions
- How the integration strategy remains consistent across all grid levels
- Different load‑balancing approaches for nonuniform workloads
- Practical notes on parallelization, including domain partitioning by level and update synchronization
Ideal for readers seeking a concrete, algorithm‑driven view of AMR in parallel computing contexts.