hexfarm.rutgers.edu

Cluster Status:

All nodes are dual cpu's. Nodes 8 and above have AMD CPU's that are 80% faster than the CPU's in nodes 2-7 and in hexfm1, hexfs1, and hexcaf.
hexcaf runs on all nodes (multi-sectioning batch facility).
Status history

Introduction:

Connection to the farm is via "ssh" to "hexfarm.rutgers.edu" or (equivalently) "hexfm1.rutgers.edu". hexfm1 is intended for debugging, not long duration runs.

There are 23 nodes for running production jobs (both non-caf and caf style). These are "node1" through "node23". For non-caf use, connect to these via ssh from hexfm1.rutgers.edu. Use the cafgui for caf jobs.

hexfs1.rutgers.edu is used for import/export of data.

hexcaf.rutgers.edu is used as the head node for CAF-RU.

Hexfm1 has 2GB of memory, all other nodes have 1GB of memory. Hexfm1, hexfs1, hexcaf, and node2 through node7 have dual 1GHz Pentim III cpu's. Nodes 8 - 15 have dual 1.5 GHz AMD cpu's. Others have dual 2.0 GHz AMD cpu's.

Data disks total about 5 TB with 3.3 TB managed by the Disk Inventory Manager (DIM) and the rest in scratch disk (or static data disk) /scr1 through /scr7.

There are two tape drives on hexfm1:

Here are the results of the CDF "tiny" benchmark run on these nodes.

User Hints:

Data in farm:

Monitoring links:

Real time system status:

Other: