SLab:Systems
Contents
Schuster Lab Computing Resources
linne cluster
The linne cluster runs SGE 6.0u6 and can be used by anyone to submit arbitrary jobs. In addition, the linne cluster runs BioTeam iNquiry, which can be accessed here.
- head node: linne.bx.psu.edu
- Apple Xserve G5
- Two 2.3 GHz PowerPC G5 processors
- 6GB RAM
- 75GB /scratch
- MacOS X 10.4
- 31 compute nodes: node001.linne.bx.psu.edu --> node031.linne.bx.psu.edu
- Apple Xserve G5
- Two 2.3 GHz PowerPC G5 processors
- 2GB RAM
- 75GB /scratch
- MacOS X 10.4
- 5 compute nodes: node036.linne.bx.psu.edu --> node040.linne.bx.psu.edu
- Dell PowerEdge 1950
- Two quad-core 2.83 GHz Xeon E5440 processors (a total of 8 cores)
- 16GB RAM
- 100GB /scratch
- Red Hat Enterprise Linux 4.5
persephone cluster
The persephone cluster is a member of the central BX SGE installation. Queue status can be seen at http://qstat.bx.psu.edu
Physical specifications can be seen in the Ganglia Physical View for Persephone: http://ganglia.bx.psu.edu/?p=2&c=persephone
There is one global all.q, with appropriate access restrictions to limit jobs to the appropriate nodes. It is absolutely essential that you specify your job's exact requirements when submitting with qsub (memory, # of cpus, etc). c1 and c2 are in a special 454pipeline.q, used for off-rig analysis of 454 Titanium runs, which only the FLX sequencer users can submit to (schuster-flx[1234]).
s1 (s1.persephone.bx.psu.edu and s1.linne.bx.psu.edu)
s1 is dual-homed, it is connected to both the persephone and linne network switches. s1 can be used by anyone to run arbitrary jobs. It is especially useful for jobs that need a lot of local disk space.
- HP ProLiant DL380 (G5)
- Two quad-core 3.0 GHz Xeon X5450 processors (a total of 8 cores)
- 16GB Ram
- 3TB /scratch
- CentOS 5.5
