Chapter 5. Using GPUs to Run R Even Faster

In this chapter, we will look at another means to speed up the execution of an R code using a technology that is often untapped, although it is part of most computers—the Graphics Processing Unit (GPU), otherwise known as a graphics card. When we think of a GPU, we often think of the amazing graphics it can produce. In fact, GPUs are powered by technologies with highly parallel processing capabilities that are like the top supercomputers in the world. In the past, programming with GPUs was very difficult. However, in the last few years, this barrier has been removed with GPU programming platforms like CUDA and OpenCL that make programming with GPUs accessible for many programmers. Better still, the R community has developed a few packages for R users to leverage the computing power of GPUs.

To run the examples in this chapter, you will need an NVIDIA GPU with CUDA capabilities.

This chapter covers:

• General purpose computing on GPUs...