Service DescriptionITS Research Computing manages a variety of servers and compute clusters to help meet the computational needs for Carolina researchers. A wide variety of commercial and open source software applications is provided for use, as are multiple compilers and parallel operating environments. These systems, located in and administered by Research Computing, currently include: i. Topsail cluster – Linux based, >4000 cores, IB interconnect, better suited for parallel than serial work. 12 GB/memory per 8 core blade. ii. Emerald cluster – Linux based, >900 cores, GigE interconnect, lots of commercial software, some large memory nodes, good for serial or small parallel runs. Using the patron or “condo” model, faculty can purchase additional compute nodes for this environment, with priority use of those nodes but making the nodes available to the community when they would otherwise be idle. iii. Kure cluster – Soon to be production; available now for friendly users. More than 450 cores, 4xQDR IB interconnect, all nodes have 8 cores and a minimum of 48 GB of memory. Good for parallel jobs and large memory jobs. Using the patron or “condo” model, faculty can purchase additional compute nodes for this environment, with priority use of those nodes but making the nodes available to the community when they would otherwise be idle. iv. Windows-based computing through the VCL (Virtual Computing Lab), providing “any time, anywhere” access to a wide range of Windows-based research software applications. Windows XP, Vista and Windows 7 environments are available. v. Linux sandbox environments– also based on the VCL (Virtual Computing Lab), providing “any time, anywhere” access to base Linux environments upon which you can build a customized Linux environment for your research needs. Red Hat, CentOS, and Ubuntu base environments are available.
Hours of AvailabilityServices are available 24x7 other than during planned maintenance periods or unanticipated emergencies (such as loss of commercial power to the data center). As the compute clusters comprise large numbers of pieces of commodity hardware, at any point in time, an individual server may not be available. Failures of individual servers outside business hours will be investigated during the next business day. However, other servers typically will be available.
Faculty, postdocs, staff and graduate students engaged in research. Undergraduate accounts for a course or specific research project may be provided with a supporting letter from a faculty member.