welcome

Thank you for purchasing the MEAP edition of Grokking Parallel Programming: With examples in CUDA.

The world’s compute has moved to the GPU. Every large language model, every image generator, and most of modern scientific computing runs on chips that execute tens of thousands of threads at once. Yet for most developers the GPU is still a black box hidden behind a library call like tensor.cuda(). This book opens the box. By the end, you will not just call parallel code; you will write it, debug it, and understand why it is fast or slow.

I wrote this book for readers who can write basic C code and nothing more. You need no prior experience with parallel programming, graphics, or GPUs. The book teaches in the Grokking style: every algorithm is unfolded step by step in carefully designed diagrams before any code appears, and every chapter ends with a Common Mistakes section built from the bugs beginners actually write.