80 likes | 268 Views
Book Review. CUDA PROGRAMMING: A DEVELOPERS GUIDE TO PARALLEL COMPUTING WITH GPUS SHANE COOK. Categories of CUDA Topics. November 2012 576 Pages $48.70. Table of Contents:. A Short History of Supercomputing Understanding Parallelism with GPUs CUDA Hardware Overview Setting up CUDA
E N D
Book Review CUDA PROGRAMMING: A DEVELOPERS GUIDE TO PARALLEL COMPUTING WITH GPUS SHANE COOK
November 2012 576 Pages $48.70
Table of Contents: • A Short History of Supercomputing • Understanding Parallelism with GPUs • CUDA Hardware Overview • Setting up CUDA • Grids, Blocks and Threads • Memory Handling with CUDA • Using CUDA in Practice • Multi-CPU and Multi-GPU Solutions • Optimizing Your Application • Libraries and SDK • Designing GPU-Based Systems • Common Problems, Causes, and Solutions
Good Impressions • Treatment of many unique topics • Good figures, diagrams • Broad, broad topic range • Breadth makes it best single introduction to CUDA
Bad Impressions • Broad, broad topic range • Inconsistent level of detail • Unedited text (minor typos and rambling) • Very poor Kepler discussion • Not a reference nor a textbook nor an overview
Complementing references: REFERENCE: NVIDIA CUDA Toolkit Documentation CLASS TEXT: Programming Massively Parallel Processors (Kirk and Hwu) BROAD CODING INTRODUCTION: CUDA Application Design and Development (Farber) CUDA ALGORITHMS AND EXAMPLES: GPU Computing Gems series