- There are three for statements in this chapter's Mandelbrot example; however, we can only parallelize over the first two. Why can't we parallelize over all of the for loops here?
- What is something that Amdahl's Law doesn't account for when we apply it to offloading a serial CPU algorithm to a GPU?
- Suppose that you gain exclusive access to three new top-secret GPUs that are the same in all respects, except for core counts—the first has 131,072 cores, the second has 262,144 cores, and the third has 524,288 cores. If you parallelize and offload the Mandelbrot example onto these GPUs (which generates a 512 x 512 pixel image), will there be a difference in computation time between the first and second GPU? How about between the second and third GPU?
- Can you think of any problems with designating certain algorithms or blocks of code as parallelizable in the context of Amdahl's Law?
- Why should we use profilers instead of just using Python's time function?
-
Book Overview & Buying
-
Table Of Contents
Hands-On GPU Programming with Python and CUDA
By :
Hands-On GPU Programming with Python and CUDA
By:
Overview of this book
Hands-On GPU Programming with Python and CUDA hits the ground running: you’ll start by learning how to apply Amdahl’s Law, use a code profiler to identify bottlenecks in your Python code, and set up an appropriate GPU programming environment. You’ll then see how to “query” the GPU’s features and copy arrays of data to and from the GPU’s own memory.
As you make your way through the book, you’ll launch code directly onto the GPU and write full blown GPU kernels and device functions in CUDA C. You’ll get to grips with profiling GPU code effectively and fully test and debug your code using Nsight IDE. Next, you’ll explore some of the more well-known NVIDIA libraries, such as cuFFT and cuBLAS.
With a solid background in place, you will now apply your new-found knowledge to develop your very own GPU-based deep neural network from scratch. You’ll then explore advanced topics, such as warp shuffling, dynamic parallelism, and PTX assembly. In the final chapter, you’ll see some topics and applications related to GPU programming that you may wish to pursue, including AI, graphics, and blockchain.
By the end of this book, you will be able to apply GPU programming to problems related to data science and high-performance computing.
Table of Contents (15 chapters)
Preface
Free Chapter
Why GPU Programming?
Setting Up Your GPU Programming Environment
Getting Started with PyCUDA
Kernels, Threads, Blocks, and Grids
Streams, Events, Contexts, and Concurrency
Debugging and Profiling Your CUDA Code
Using the CUDA Libraries with Scikit-CUDA
The CUDA Device Function Libraries and Thrust
Implementation of a Deep Neural Network
Working with Compiled GPU Code
Performance Optimization in CUDA
Where to Go from Here
Assessment
Other Books You May Enjoy
