CCBM Computational Resources

Hardware:

1 IBM BladeCenter cluster composed of 1 IBM x345 head node with 2 2.8GHz processors and 2.5GB memory; and 63 HS20 blades, each with 2 2.8GHz processors and 4GB memory, for a total cluster core count of 128. The cluster uses the Rocks Clusters 4.1 cluster operating system, including Sun Grid Engine; LAM/MPI, and the Intel compiler suite.

2 Dell PowerEdge R900 servers; one with 2 Quad Core 2.4GHz processors and 64GB memory, the second with 4 Quad Core 2.4GHz processors and 128GB memory; both running CentOS Linux.

5 IBM xSeries servers, and 2 Dell PowerEdge 2950 servers, each with Dual Core Intel 2.66GHz processors and 2GB RAM.

1 IBM p690 with 32 1.1GHz Power4 processors, and 128GB of memory, running IBM AIX 5.2.

10TB direct-attached storage connected to 1 Dell R900. 6.8TB of storage in an IBM FAStT500 SAN.

1 IBM 3584 Ultrascalable Tape Library with 4 LTO drives (soon to be upgraded to 4 LTO4 drives). Tivoli Storage Manager 5.5 is used to manage backup, archive, and restore.

CSEB 213 machine room:

The machine room is located on the 2nd floor of the new Computational Sciences and Engineering Building on the Homewood campus of JHU. The room is 1000 sq. ft., with space for 22 standard 42" x 24" racks in 3 aisles. Power is supplied by a dedicated 300kVA transformer. Cooling is provided by two 25-ton computer room air-conditioning units, with facilities and space available for a third. This equipment is managed and monitored by JHU Plant Operations 24/7 and is configured to generate audible and visual alarms as well as reporting back to JHU Plant Operations. Additional environmental monitoring tools are in place to send alerts to local systems administration staff.

The new Computational Sciences and Engineering Building has a 10GbE backbone, providing 10GbE connectivity to the JHU Homewood campus core, which is one hop away. The CSEB 213 machine room has a dedicated Cisco 4500 series switch, which will provide 10GbE connectivity to the equipment housed there. All networking equipment is managed and monitored by Johns Hopkins Enterprise