ABSTRACT
This chapter provides insight into the implementation, benefits and problems, current solutions, and future challenges of systems having CPUs and GPUs on the same chip. It describes two modern processors, AMD Llano and Intel Ivy Bridge, which have both integrated central processing units (CPUs) and graphics processing units (GPUs) on the same die. These chips are often referred to as accelerated processing units (APUs). The integration of CPU and GPU on the same die will help ease memory bandwidth constraints. With Moore's law, on-die integration of many components such as peripheral control hubs, dynamic random access memory dynamic random access memory (DRAM) controllers, modems, and more importantly, graphics processors have become possible. Single-chip integration of GPUs with CPUs has emerged and also brought many challenges that arise from integrating disparate devices/architectures, starting from overall system architecture, software tools, programming and memory models, interconnect design, power and performance, transistor requirements, and process-related constraints.
