Programming Massively Parallel Processors: A Hands-on Approach

The new edition of the book is improved significantly related to the already superb previous edition.The technical writing and presentation is outstanding and a lot of CUDA programming examples are presented.Also, it covers thoroughly the recent advances in CUDA technology as the dynamic parallelism.I strongly recommend the book to the Amazon readers interested in GPU computing.

Well written and very useful, this is the best book to start learning CUDA programming from.

Thorough coverage of the hardware, data structures, and algorithms you need to effectively use Nvidia GPU devices. Missed one star, because the code examples are incomplete: kernels for GPU execution are well documented, but the host code is missing, so outside resources (books, websites, etc.) are required to write compilable programs. Please, open your source code!

This book (third edition) pretends that you can go online for supplements, like any other programming book, but it's just kidding. There's a course page with nothing on it but a course description. There's a lab page that actually lets you download a zip file with completely empty folders in it. There's zero assistance in the book with setting up a development environment, even though it's apparently difficult enough for the course description to dedicate a whole lab to it. And most problematic, there won't be any files of data big enough to use as input to the programs to actually demonstrate massively parallel processing. I'm quite sure that the only programs you'll be able to run using just the book, in every chapter, will barely rise above hello world.

Perfect condition, fast shipping!

This book is a lot better than others I have read.

Only CUDA programming book that contains Pascal Arch. right now.

Excellent product.

Having some C++ experience, I was looking for a comprehensive introduction to CUDA and its optimisations. This book is exactly what I was hoping for and gave me all the information I was struggling to find online. The parts about performance optimisation are so helpful and I was able to finish the project I was working on thanks to this book.I would say, as one other review noted, only the device code is given, meaning it's sometimes not easy to intuit the host code which is left entirely to the reader. I understand constraints of space in the book, but I don't think this was the best approach.For that reason, I don't recommend this to complete beginners. I don't think I would have been able to do this without some existing C/C++ experience.

Para mí es uno de los mejores libros, y en el que me baso para ir dando mis primeros pasitos en CUDA. Como valor añadido puedo deciros que al haber conocido a los autores del libro en una conferencia, ambos han realizado un libro realmente impresionante y decir que tienen muy en cuenta las opiniones constructivas que se les dan.

The text is great. Easy to understand, concise, illustrated with examples and execercises. However, the quality of the book itself is miserable. You must not bend it, neither open it further then 60 deg. Otherwise, the glue breaks and a hundred pages flip into you hands.