High-Performance Computing

High-Performance Computing is a modern guide to the architecture, software, mathematics, and engineering principles that determine how fast scientific computation can actually run.

Rather than treating performance as a matter of processor speed alone, the book shows how time to solution emerges from the interaction between algorithms, microarchitecture, memory, storage, networks, parallelism, accelerators, numerical methods, and software design.

Across twenty-one chapters, it develops the foundations of performance analysis before moving through OpenMP, MPI, GPUs, heterogeneous computing, linear algebra, scientific simulation, cluster architecture, schedulers, fault tolerance, cloud HPC, artificial intelligence, and post-exascale systems.

A central theme is measurement before optimisation. Peak specifications are constantly separated from achieved performance, and equations are derived rather than merely quoted. Worked examples and exercises connect the theory directly to practical performance engineering.

Written for students, researchers, engineers, and computational scientists, High-Performance Computing provides a rigorous framework for understanding what limits a computation, how to measure it honestly, and what must change before a faster result can truly be claimed.