“C++ AMP: Accelerated Massive Parallelism with Microsoft Visual C++” is a good introduction to programming with C++ AMP.
Unfortunately, at the time of this review, it is the only book available on C++ AMP, so I can only
compare it with other books on CUDA and OpenCL.
But the book stands up to the comparison and has earned a solid 4 stars.
I found the order in Chapter 2 to be a bit awkward, but that settled down in the next chapter.
You should skip Chapter 2 during the first reading and only read it at the end.
This is because Chapter 2 covers the somewhat more complicated n-body problem and even includes DirectX graphics.
From a pedagogical standpoint, it is not wise to deal with this at the beginning.
Then there are a few minor details, such as the fact that std::vector<> and std::shared_ptr<> are constantly written out in the program code.
The std:: bothers me.
But all in all, after reading this book, you are already an advanced GPU programmer
who has a good understanding of the subject. Prior knowledge of C++, STL, and parallel programming is required.
The example programs have been adapted for VS 13.
- Kate Gregory, Ade Miller
- C++ AMP: Accelerated Massive Parallelism with Microsoft Visual C++
- Microsoft Press
- 2012
See also the review on Amazon.
