A massively parallel computing system which will be dedicated to support research in computer science will be purchased. The equipment will be used for several research projects including in particular: Signal processing and data analysis employing fast Fourier transform algorithms for non-abelian groups and homogeneous spaces and wavelet and other phase space techniques. Experimentation with parallel implementation is an important new direction for this work. Applications in medical imaging are emerging from this effort. Building a prototype checkpointing system for multiprocessor file systems on massively parallel computers, with follow-on-research in interfaces, caching and prefetching, workload studies, and remote heterogeneous file system access. Development of virtual memory for data-parallel computing with applications in scientific computing. Applications in image processing involving ongoing projects in human brain labeling, fossil leaf classification, handwriting analysis, and simulation of visual attention. Implementation of PRAM algorithms aimed at elimination of concurrent references to memory in implementations on a massively parallel synchronous computer, using techniques taken from work of the current researchers in new algorithms for the PRAM model.
