PSI Project HP

Introduction to the i3 Cluster

The I3 Cluster currently has 16 compute nodes; i{1-15}.Millennium.Berkeley.EDU
  • 16 HP Proliant DL140 (specs)
    • Dual 3.06GHz Pentium 4 Xeon (533Mhz FSB)
    • 512K L2 cache
    • 2GB of RAM
    • 2 80GB 7500 rpm IDE disks
    • Broadcom 1000TX gigabit Ethernet
  • 1 HP Procurve 2848 Gigabit Switch

Jul 2006: i6 is offline due to hardware failure.
Apr 2007: i16 has been renamed i6.
Mon Apr 23 2007: The i3 cluster will be in use for undergraduate instruction until the end of the semester.

The IBM Fast Storage Cluster provides high-speed shared file systems to this compute cluster.

Using the Cluster

Access to this cluster is limited to members of the i3 Project.  If you are an i3 member, submit an account request.

This is an unscheduled shared resource. You can view the current use of the machines by running gstat or viewing the Ganglia graphs.

gstat provides an ordered list of available machines. Machines with a load >= 2 are fully loaded and should be avoided until current jobs have completed.

Run gexec to launch jobs.

cd /work/$user
gexec -n 10 hostname (returns the hostname of 10 machines)
cp $home/foo /work/$user
gexec -n 20 ./foo (runs foo on 20 machines)

For more information on using gexec, please see the documentation.

Community & Reservations

Use the Cluster Reservation webpage to schedule time on the cluster.  Login as "i3" using the PI's office number as the password. This is a socially enforced calendar to aid in cooperative scheduling.

The i3-users mailing list is a forum for community support as well as formal announcements regarding the i3 cluster.

Filesystems

While you should be able execute jobs from your EECS department home directory, we strongly suggest that you launch all jobs from a /work/$user or /scratch/$user directory. Before executing a program, copy all binaries and data files into your /work or /scratch directory, cd into that directory, and execute from there. This avoids putting unnecessary load on EECS department fileservers, which are sometimes unable to handle many simultaneous mount requests.

Note: /work has a 30 day deletion policy. Files in /work untouched for 30 days will be purged without warning. /work is not for long-term storage and is never backed up. /work is shared with the PSI cluster.

/scratch has NO deletion policy; please clean up after yourself. /scratch is RAID 0 local to each machine. /scratch is never backed up.

UC Berkeley Clustered Computing - Last modified on 25-Apr-2007 16:45:14 -0700