High Performance Computing
Parallel Computing
Parallel computing or high performance computing utilizes multiprocesses to perfrom concurent jobs in parallel. The technique can help speedup computation. Professional staff at ITS AT-Research Service (RS) have more than two-decades experience of high performance computing and parallel computing. ITS has been providing high-quality support to UI faculty who are using parallel computing technology to solve their research problems. Currently, RS maintains and administrates a couple of LINUX clusters for public use as well as several LINUX clusters for faculty use. RS staff collaborates with UI faculty for pursuing research grants and help departmental research groups to establish HPC labs. In general, RS's services includes HPC hardware support, system maintenance and administration, parallel software installation, parallel system configuration and consultation, parallel algorithm design and programming development, and training through credit courses and workshops.
Parallel Programming
Parallel programming plays an important rule in parallel computing. It focuses on parallel algorithm development, and parallel utility or library implementation. ITS RS staff are enthusiastic in helping UI researchers to design, program, and implement parallel jobs on HPC resources. RS staff help UI researchers to apply for allocation time from national supercomputing centers. RS staff provide client-oriented services of MPI-enhanced parallel programming, as well as other parallel utility libraries on LINUX clusters.
Parallel Distributed Computing
Parallel distributed computing emphasizes parallelism of computations over network-based systems. ITS-RS has great expeirence in such computations. RS staff can work closely with UI researchers to conduct such computations on the internet by reducing network latency, balancing job-loading, and testing performance of networking data transfer and capacity of computing at remote systems.
Grid-enhanced HPC
Grid-enhanced HPC is high-end parallel distributed computing. It is based on "virtual organization" to conduct large scale, intensive computations crossing different institutions. At this stage, RS staff master many Grid technologies during many Grid projects, enveloped in GROW projects to promote Grid computing applications and research. At this stage, RS staff work closely with other UI faculty to do research in Grid computing, and transfer the Grid technology to UI faculty.
Contact
For further information on any of these topics, please contact:
Jeff DeReus
jeffrey-dereus@uiowa.edu
