The Parallel & Distributed Systems (PDS) Group is part of the High Performance Computing Laboratory (HPCLab). PDS Group includes faculty members and graduate students, of the Computer Engineering & Informatics Department, conducting research in various topics of parallel and high performance computing.
Our group activities concern research conducted on systems software and middleware architectures for parallel and distributed computing. We investigate fundamental algorithms and heuristics for parallel processing, we study and design protocols that optimize the use of deep memory hierarchies, and we also concentrate on inter-process and inter-node communication. Respectively, we focus on scheduling policies and the support of efficient multithreading by designing and developing user-level as well as kernel-level software environments and tools, that enable and enhance high performance computing on homogeneous and heterogeneous parallel platforms.
The current research activities include efficient runtime systems for shared memory programming models, software distributed shared memory (S-DSM) architectures, and operating system scheduling and resource allocation algorithms for multi-core and multiprocessor systems. Our group research efforts also include environments for Grid Computing and the design of a resource management framework for the dynamic formation of Computational Grids throughout the Internet.
The PDS Group is lead under the supervision of E. Polychronopoulos faculty member of the Computer Engineering & Informatics Department.