Scientific Computing in Modern C

If you are an aspiring computer programmer with a basic knowledge of C++ and want to deepen your understanding of the language, specifically for use in scientific computing, this unique book is for you. It is the only book to discuss modern C++ from a scientific computing perspective, covering a wide range of topics¬¬, including the finer points of C++, specific idioms of C++ in scientific computing, parallelism, considerations of hardware and performance, and “carpentry” topics, such as CMake, that extend beyond basic programming to make you a more productive programmer. This book focuses on computing “idioms” and applications—rather than a complete treatment of the C++ language—as well as peripheral “carpentry” topics, and is designed to take students to a level where they can be productive in a scientific programming project. The C++ topics discussed are chosen for their relevance to computing, and other topics are purposely excluded. Additionally, several topics relevant to scientific computing are included that are not intimately tied to C++ as a language.

Victor Eijkhout was brought up as a numerical analyst, coding in Fortran and proving theorems about the asymptotic behavior of parallel numerical methods. These days he mostly writes in C++, both sequential and parallel in various ways. While he no longer proves many theorems, he is still interested in code performance for finite numbers of processors. He joined the Texas Advanced Computing Center in 2005, where he supports research, teaches various aspects of high-performance computing, and flexes his programming muscles on large supercomputers.