None
The text discusses the challenges of parallel programming and the difficulties programmers face when writing explicitly parallel programs. The author argues that parallel programming is performance-oriented, meaning that the program must not only be correct but also fast. To achieve this speed, programmers must divide applications into smaller parts to be processed by multiple processors, while minimizing scheduling and coordination overhead. The analogy of eight reporters working on a story highlights the challenges of parallel programming, including load balancing, synchronization, communication overhead, and managing dependencies between tasks. The text also introduces various approaches to parallel programming, including shared addressing and explicit message passing, and discusses the difficulties of benchmarking parallel processors. Finally, the text mentions that graphics processing units (GPUs) are becoming popular programming platforms due to their parallel architecture.