Covers multi-core GPU programming with some emphasis on CPU vs GPU hardware, with benchmark/performance tests to show improvements from purely a CPU approach to a GPU approach to processing data using several different libraries including CUDA, OpenCL, and DirectCompute.