Chapter 11. CUDA for Real Problems
CUDA gives developers the ability to accelerate applications on the GPU by one to three orders of magnitude over conventional processors. Knowing how to think and program in CUDA is the necessary first step in the process of growing and maturing into an adept CUDA programmer. Look to the Internet and technical literature to find projects that provide general solutions and design patterns that can be adapted to your projects. CUDA is fast; reinventing the wheel is slow. The chapters in this book were designed to teach and provide some initial frameworks for future applications. For example, the simple C++ functor framework in Chapter 2 implemented a general parallel mapping for pattern recognition and optimization ...
Get CUDA Application Design and Development now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
