Programming Massively Parallel Processors: A Hands-on Approach, Fifth Edition shows both students and professionals alike the basic concepts of parallel programming and GPU architecture. Concise, intuitive, and practical, it is based on years of road-testing in the authors' own parallel computing courses. Various techniques for constructing and optimizing parallel programs are explored in detail, while case studies demonstrate the development process, which begins with computational thinking and ends with effective and efficient parallel programs. This new edition has been updated with an expanded repertoire of optimizations, new patterns and applications, ad more coverage of important CUDA features. - Expanded optimization checklist with a more comprehensive demonstration of essential optimizations across patterns- New pattern and application chapters including: filtering, wavefront parallelism, advanced optimizations for matrix multiplication, and large language models (LLMs)- More coverage of important CUDA features including warp-level programming, cooperative groups, CUDA C++ atomics, and multi-GPU programming with NCCL and NVSHMEM

Programming Massively Parallel Processors: A Hands-on Approach, Fifth Edition shows both students and professionals alike the basic concepts of parallel programming and GPU architecture. Concise, intuitive, and practical, it is based on years of road-testing in the authors' own parallel computing courses. Various techniques for constructing and optimizing parallel programs are explored in detail, while case studies demonstrate the development process, which begins with computational thinking and ends with effective and efficient parallel programs. This new edition has been updated with an expanded repertoire of optimizations, new patterns and applications, ad more coverage of important CUDA features. - Expanded optimization checklist with a more comprehensive demonstration of essential optimizations across patterns- New pattern and application chapters including: filtering, wavefront parallelism, advanced optimizations for matrix multiplication, and large language models (LLMs)- More coverage of important CUDA features including warp-level programming, cooperative groups, CUDA C++ atomics, and multi-GPU programming with NCCL and NVSHMEM