The College of Staten Island of the City University of New York (CUNY) plans to add a new Cray system, code-named Cray-Baker, to its existing set of machines to exploit the advantages provided by Partitioned Global Address Space (PGAS) programming languages, such as Co- Array Fortran and Unified Parallel C. The significance of this project is that it allows CUNY and collaborating researchers to jump-start academic research in the development of PGAS-based applications by acquiring the first such system, outside of the National Security Agency and Department of Energy. City University of New York (CUNY) investigators use high performance computing (HPC) to simulate electro-mechanical coupling in the heart to understand how acquired disorders, such as myocardial infarction, and genetic diseases, result in abnormal heart rhythms. CUNY investigators also use HPC to model the ability of urban structures to withstand the effects of blast from explosive devices. The new HPC system with its advanced programming models and support for low-latency global communications accelerate these research activities. Collaborations with noted national and international researchers are supported.
The award by the National Science Foundadtion (NSF) of Major Research Instrumentation Grant 0958379 to the College of Staten Island of the City University of New York (CUNY) was instrumental in introducing a new community of researchers and students to High-performance Computing (HPC) and computational science. CUNY consists of a population of 270,000 students in degree programs, 220,000 students in non-degree programs, and 7,300 faculty members. We do not claim that a large percentage of faculty and students are either aware or use these resources. We do claim, however, that this MRI award has significantly raised community awareness of HPC and computational science and has also been instrumental in increasing the number of users of HPC at CUNY. The MRI funded the acquisition of a Cray XE6m with 1,280 compute. The impact of this instrumentation was to facilitate increased use of HPC in diverse areas of computational science; (2) encourage the increased use of new programming models such as Coarray Fortran (CAF) and Unified Parallel C (UPC); and, (3) encourage faculty and students to become more aware of the importance of high-performance computing and to incorporate computational methods in research and classwork. The MRI award also allowed CUNY researchers to apply for and receive new research grants and also increase collaborations with researchers at other institutions. A year after the installation of the NSF-funded Cray XE6m, CUNY itself upgraded the system by more than doubling the number of compute cores from 1,280 to 2,816. In addition, research collaborators in the Consortium for Advanced Research Transport of Hydrocarbon in the Environment at the Rosenstiel School of Marine and Atmospheric Science, University of Miami, based on their experience on the CUNY Cray XE6m, procured their own system to expand their own resources. The City of New York, in June 2013, in the wake of Hurricane Sandy, issued a report titled, "A Stronger, More Resilient New York". The widely acclaimed report is a plan to enhance the City’s infrastructure to make the City more resilient to future storms. Many of the computational studies of risk assessment used the ADCIRC/SWAN model and were run on the NSF-funded Cray XE6m by the CUNY HPC Center and researchers from the Renaissance Computing Institute. Researchers at the CUNY Cooperative Remote Sensing Science and Technology (CREST) Institute developed a fine scale "urban" Weather Research and Forecasting Model (uWRF). uWRF gives additional capability to the standard WRF Model by capturing the influence of urban areas on wind, temperature, and humidity in the boundary layer through a multilayer urban parameterization and energy model that accounts for impacts from horizontal and vertical building surfaces in the conservation and turbulence equations considering radiation exchange between indoor surfaces and generation of heat due to occupants and equipment. "Our modeling takes into account unique characteristics of urban areas and enhances our ability to accurately anticipate extreme weather events for densely populated urban areas. We have configured this modeling approach and tested under heat wave conditions for the City of New York with very optimistic results", said Dr. Jorge Gonzalez, CREST Professor of Mechanical Engineering and project Principal Investigator. The effort was a collaboration between the CREST Institute, NCAR, San Jose State University, the NSF supported CUNY-High Performance Computer Center, and the Spanish Research Center for Energy, Environment and Technology (See Figure 1). In addition to enhancing city and regional near-term weather forecasts and storm prediction, uWRF represents a valuable tool to understand the larger implications of the interrelated natural-human system. The instrumentation has become an important resource in the academic environment. In the 2012-13 Academic Year, for example, eight undergraduate and graduate credit courses in high performance computing were conducted using the resources of the CUNY HPC Center including the instrumentation. Training classes were held in the use of Coarray Fortran and Unified Parallel C. A number of researchers in environmental science and molecular dynamics quickly adopted these programming models as a more efficient method for developing scalable applications, although the limited number of computer platforms that effectively support these efficient programming models remains an impediment to their wider adoption. Since the installation of the instrumentation, based upon 24x7 operating schedule, uptime has been greater than 98% and utilization in excess of 90%.
