Abstract

The CUNY-HPCC was created in 2008 and quickly become core research facility at the University. The center supports mixture of architectures (SMP, DSM, Clusters) all integrated together and thus supporting wide variety of research activities. The center implements advanced  job management system (beyond basic batch system) systems which analyzes jobs and places them optimally in terms of resource utilization and jobs throughput.  In this presentation we will discuss the High Productivity Computing (HPrC) as it is implemented at CUNY-HPCC especially the support for multi scale simulations.  Indeed, most mathematical operations do have some degree of parallelism – from perfect mathematical parallelism when no dependencies among the different pairs of data exist, to partial parallelism which incorporates interdependencies. On other hand the modern HPC systems offer several levels of parallelism from CPU based parallelism to system defined parallelism to SIMT parallelism of graphics processing units (GPUs). In that sense we will discuss the different programming models supported by CUNY-HPCC, tools, job and data management, application management and user support and analytics as core elements of HPrC at CUNY.

Bio

Dr. Tzanov is the Director of CUNY High Performance Computing Center at College of Staten Island. His background is in Applied Physics and Electronics, Environmental Science, Computer Science and Computational Chemistry. He worked as tenured faculty at Departments of Computer Science  at Technical University of Sofia and at Queens University of Belfast (UK) and as visiting professor at Buckingham University (UK). He is also assistant professor of professional practice at Rutgers University. He did his post doc with Prof. Barry Honig at Columbia University (1998-2001) working on developing novel methods for protein structure prediction. His PhD is in Computational Chemistry at New York University. He worked at NYU HPC division and at Department of Chemistry at New York University for more than 13 years before moving to CUNY-HPCC in 2013. His theoretical research interest is in development of enhanced sampling methodologies for the field of Computational Chemistry and Materials Science, AI in Chemistry and Physics, Scientific High Performance Computing and scalable numerical algorithms.