Irregular computations on heterogeneous multi-core architectures

点击翻译按钮获取中文摘要

Parallel computing involves using many computers to solve a given problem fast and cooperatively. Problems in medicine, transportation networks and finance produce significant amounts of data that require processing of the information in real-time. Parallel computing becomes fundamental for these applications. These days, general-purpose computers (PC) come with not one CPU (central processing unit or processor) but with 2, 4 or 8 identical processors (or cores), allowing simultaneous execution of many tasks. Recently, graphical processing units (GPUs) have become mainstream in applications such as games, video, and multimedia. There are more than hundreds of processing elements on a single GPU providing lots of data parallelism (for example, each pixel on the video screen can be processed independently). These GPUs come as a single chip and can be installed on any PC. CPU together with GPUs provide massive amount of parallelism and are called multi-core machines. Play Station (PS3) is a typical example of a multi-core machine. The question one asks is what are we going to do with these powerful computers and what applications besides games can be programmed on these machines? One of the applications this research concentrates on is medical imaging. Determining the internal structure of an organism, with sufficient detail to yield diagnostic information of a patient, is a first step before administering any procedure. Medical imaging is the process by which physicians evaluate an area of the patient's body that is not normally visible. There are many imaging techniques. This research focuses on computed tomography and emerging microwave tomography techniques for breast cancer detection. The images in any of the imaging techniques require large amounts of storage space and need to be processed fast to make real-time decisions on the patient. Therefore, the focus of this proposed research will be on the design and performance evaluation of sophisticated algorithms on multi-cores for problems in medical imaging techniques. This research will also focus on computational issues in using the powerful computers for problems in transportation network and finance.