City University of New York (CUNY) researchers use high performance computing (HPC) to study air pollution in urban environments, develop techniques for visualization of scientific data, develop Monte Carlo-based algorithms for modeling of complex physical and transportation systems, and, with collaborators at the New York State Institute for Basic Research in Behavioral Disabilities (IBR), develop computational methods for autism studies. The investigators use a unique100-teraflops parallel heterogeneous HPC system, funded by NSF, which consists of 96 Intel 5570 quad-core processors connected to 96 NVidia Tesla graphics processing units. The Heterogeneous High Performance Computing System (H-HPCS), is tailor made for computations that are numerically intensive, vectorizable and highly parallel.
CUNY scientists use H-HPCS to process atmospheric data obtained from satellites, lidar, and ground weather stations to assess pollution concentration and dispersion. Energy production and vehicular traffic within the City of New York generate large amounts of chemical and fine particulate pollution, which can induce or have adverse affects on residents, particularly those with cardio-respiratory illnesses. The H-HPCS techniques are developed to predict the dispersion of fixed point and mobile releases of chemicals and biological agents. IBR researchers conduct studies on the detection of autism related behavioral disorders using H-HPCS analysis of genomic data and video data obtained from motion capture systems. IBR and CUNY researchers collaborate on the development of these analysis techniques, which also have corollary applicability in the creative arts. The highly parallel nature of CUNY researcher?s models enable the use of the H-HPCS to study physical systems including mixing and transport of biological and physical properties by oceanographic flows, supersolids, atomic optical lattices, quantum phase transitions, and urban transportation systems.
