: Matrix multiplication, Fast Fourier Transform (FFT), and solving linear systems Non-numerical
He introduces (the law of diminishing returns) and Gustafson’s Law (scaled speedup) early. The "Practice" side of the book then shows exactly how these theoretical ceilings manifest in code—when a programmer adds too many locks (serialization) or uses too many message-passing steps (latency).
The book provides a rigorous introduction to thread management. It covers the theory of (simultaneous access to a shared variable) and the practical solution: mutexes (mutual exclusion locks). Quinn walks through:
Amdahl’s Law: This formula helps you understand the maximum improvement possible when only a part of a system is improved.