MAGNUM
| |
System |
Type |
Number of nodes |
Number of cores |
CPU type |
Theoretical total peak |
Total memory |
Total disk capacity |
| gardar |
HP BL280cG6 |
cluster |
288 |
3456 |
Xeon |
35 Tflop |
6912 GB |
|
| hexagon |
Cray XT4 |
MPP |
1388 |
5552 |
Opteron |
51 Tflop |
6064 GB |
288 TB |
| stallo |
HP BL 460c |
cluster |
704 |
5632 |
Xeon 2 |
60 Tflop |
12064 GB |
128 TB |
| njord |
IBM p575+ |
distributed SMP |
192 |
2976 |
Power5+ |
22.7 Tflop |
- GB |
120 TB |
| titan |
Sun X2200 |
cluster |
304 |
2432 |
Xeon/Opteron |
21.5 Tflop |
- GB |
- GB |
| vilje |
SGI Altix 8600 |
cluster |
1440 |
23040 |
Intel Sandy Bridge |
479 Tflop |
|
|
Full name: magnum.uio.no
A summary is provided below.
More details can be found on UiO's local pages.
Magnum is a shared-memory machine with 64 Itanium2 (64-bit) processors. The machine is intended for large parallel OpenMP or MPI applications that require tight coupling (fast communication) between the processes.
Typical usage of this resource is for jobs that require 4-16 CPUs.
Single CPU large memory jobs (with memory needs larger than 4 GB) and I/O-intensive jobs also run well on this machine.
NOTE: magnum was taken out of the Notur system on December 31, 2007. Questions regarding magnum should be sent to support-uio@notur.no.
| Machine profile |
Operating system |
HP-UX |
Queueing/scheduling system |
PBS/Maui |
Total space for shared temporary storage |
2 TB |
Total space for home directories |
1.5 TB |
Max addressable memory |
100 GB |
| |
| User limits |
Default size home directory |
x GB |
| |
| Job limits |
Maximum job run-time |
2400 CPU-hours |
Maximum number of CPUs per job |
16 |
Location: University of Oslo
Installed/last upgraded: Nov 2001 / Oct 2003
| 