Maths Compute Cluster load averages


This displays the current load average of each system in the Maths compute cluster and is updated every 60 seconds. For each system, three figures are shown: the first is the load averaged over the previous 60 seconds, the second figure shows the load averaged over the previous 5 minutes and the third the load averaged over the past 15 minutes.

If a system is completely idle, the load average will be zero or close to zero. Where a load average of less than 1.00 is displayed, this means that the system is lightly loaded, has spare CPU time available and is almost certainly not running any compute jobs. A load average of around 8.00 means that the system is runing one user compute job per each of the 8 CPU cores on these systems and is using all the available CPU capacity. Load averages higher than this usually mean that more than 8 jobs are being run, each vying for CPU time on a system that is already 100% fully-loaded; it is difficult to push this figure much beyond 8 unless an operating system or disk read/write fault condition exists, in which case load averages in the hundreds have been known to occur with the system appearing largely unresponsive.

If you want to run a compute job, chose a system with a zero or near zero load average if you want the job to finish quickly!


Latest load average figures as at 17:11 Sunday 29 Jan 2012...

HP DL160 server cluster (formerly the Fünf Gruppe):

macomp01: 17:05:08 up 25 days, 5:04, 1 user, load average: 2.16, 2.03, 2.01

macomp02: 17:05:17 up 90 days, 7:08, 0 users, load average: 6.05, 6.03, 6.01

macomp03: 17:05:29 up 6 days, 4:52, 0 users, load average: 5.13, 5.60, 4.13

macomp04: 17:05:43 up 38 days, 48 min, 0 users, load average: 4.00, 4.00, 3.96

macomp05: 17:05:53 up 90 days, 17:02, 0 users, load average: 2.17, 2.07, 2.01

macomp06: 17:06:03 up 186 days, 2:32, 0 users, load average: 4.02, 4.02, 4.00

macomp07: 17:06:09 up 186 days, 5:42, 0 users, load average: 0.00, 0.00, 0.00

macomp08: 17:06:16 up 186 days, 5:01, 0 users, load average: 0.00, 0.01, 0.00

HP blade systems:

mablad01: 17:07:08 up 90 days, 17:03, 1 user, load average: 6.00, 6.00, 5.96

mablad02: 17:07:18 up 90 days, 17:03, 0 users, load average: 1.16, 1.05, 1.01

mablad03: 17:07:43 up 90 days, 17:03, 0 users, load average: 1.51, 1.19, 1.07

mablad04: 17:08:11 up 7 days, 7:40, 0 users, load average: 1.95, 1.89, 2.10

mablad05: 17:08:22 up 90 days, 17:04, 1 user, load average: 2.00, 2.00, 2.00

mablad06: 17:08:46 up 90 days, 17:04, 0 users, load average: 1.70, 1.76, 1.91

mablad07: 17:09:17 up 90 days, 17:05, 0 users, load average: 2.14, 1.94, 1.96

mablad08: 17:09:44 up 90 days, 17:05, 0 users, load average: 1.37, 1.68, 1.83

mablad09: 17:10:31 up 90 days, 17:06, 0 users, load average: 2.82, 2.10, 1.93

mablad10: 17:10:42 up 90 days, 17:06, 0 users, load average: 1.00, 1.12, 1.21

Stats Linux cluster:

fallas: 17:09:07 up 114 days, 3:33, 5 users, load average: 0.00, 0.00, 0.06

festival: is down

festival: 03:24:16 up 5 days, 15:54, 0 users, load average: 0.00, 0.00, 0.00

fete: 17:09:57 up 99 days, 6:09, 0 users, load average: 1.00, 1.02, 1.05

fiesta: 17:10:05 up 109 days, 2:45, 0 users, load average: 0.00, 0.00, 0.00

fira: 17:10:13 up 75 days, 1:56, 0 users, load average: 0.02, 0.02, 0.05

hustler: 17:10:21 up 88 days, 4:32, 2 users, load average: 0.00, 0.00, 0.00