Working with PyCUDA

In the last chapter, we saw the procedure to install PyCUDA for Windows and Linux operating systems. In this chapter, we will start by developing the first PyCUDA program that displays a string on the console. It is very important to know and access the device properties of the GPU on which PyCUDA is running; the method for doing this will be discussed in detail in this chapter. We will also look at the execution of threads and blocks for a kernel in PyCUDA. The important programming concepts for any CUDA programming, such as allocating and deallocating the memory on the device, transferring data from host to device and vice versa, and the kernel call will be discussed in detail, using an example of the vector addition program. The method to measure the performance of PyCUDA programs using CUDA events and to compare it with the CPU program will also be discussed...