UCY-CompSci – Computational Sciences Laboratory
UCY-CompSci – Computational Sciences Laboratory

UCY-CompSci – The Lab

UCY-CompSci was established as a Marie Curie Transfer of Knowledge Center (TOK-DEV) aiming at the promotion of Computational Science and Engineering and is operated by the Engineering School at the University of Cyprus (UCY-Eng) under the leadership of Prof. Stavros Kassinos.

UCY-CompSci provides a multi-disciplinary research core aiming at the promotion of excellence in a range of computational science and engineering themes that support the research activities of the School of Engineering, such as Complex Fluid and Plasma Flow, Molecular Dynamics and Ab Initio MD Simulations for Nanofluidics and Plasma-Wall Interactions, Multi-Phase Flows and Pollutant Dispersion, Biomedical Fluid Dynamics, and Turbulence Theory, Modeling  and Simulation.

UCY-CompSci has been the host of several nationally and internationally funded research projects.

 

UCY-CompSci – The Lab Equipment

UCY-CompSci has two different clusters, namely Orion and Canis.

Orion and Canis serve different needs. Orion is used for running big production jobs with in-house codes that do not need frequent library updates. Canis was designed for use by the students and staff of the Center for learning and for running software that need continuous updates of libraries and compilers. In fact, Canis is instrumental for the success of the transfer of knowledge activities.

Orion is the front-end for a 64 node (128 processor) AMD Opteron cluster. Each node has dual Opteron 244 (1.8GHz) processors with 4Gb of RAM and an 80Gb hard drive. The OS on the cluster is Linux 2.4.21-281 and the interconnect is Wulfkit (node to node latency = 3.5E-6 sec, node to node bandwidth = 200 Mbytes/sec).

A total of 3.0 T of storage is available for access during simulations. Simulation results are then backed-up to a separate archival mass-storage system.

Canis is the front-end for a 20 node (40 processor) AMD Opteron cluster. Each node has dual Opteron 244 (1.8GHz) processors with 2Gb of RAM and an 80Gb hard drive. The OS on the cluster is Linux 2.6.9-22 and the interconnect is Ethernet.

Canis is designed for use by the students and staff of the Center for learning and for running software that need continuous updates of libraries and compilers.

A total of 3.0 T of storage is available for access during simulations. Simulation results are then backed-up to a separate archival mass-storage system.