This book investigates the challenges and methodologies for performing numerical linear algebra computations efficiently on modern high-performance computers. It addresses the fundamental problem of designing algorithms that can leverage the parallel architectures and complex memory hierarchies of supercomputers to achieve optimal performance. The text focuses on the theoretical underpinnings and practical implementation of these algorithms, drawing on the extensive experience of its author in the field of high-performance computing and numerical analysis.

The book is recognized as a foundational text in the field of high-performance numerical linear algebra. Its author, J. J. Dongarra, is a leading figure whose work has significantly shaped the development of scientific computing. The text is valued for its rigorous approach to algorithm design and its practical insights into achieving computational efficiency on parallel systems. It serves as a critical resource for researchers and practitioners seeking to understand and implement advanced numerical methods.