This is an old revision of the document!
SoT cluster - the School of Technology computing cluster
Smallvoice, uses Slurm workload manager to create a computing cluster
When logged on to the cluster user is always on the login node, called freedom and should do all his work there.
/home are hosted on a NFS server, so every nodes have the same “physical” disks
All user-jobs should run using slurm sbatch job.sh, please do not run job locally on the login node
The computing (slurm) environment
There are 4 partitions / queue's and access is dependant on your account
| Name | Nodes | GPU | Timelimit | Usage |
|---|---|---|---|---|
| doTrain | 3 | Nvidia A100 GPU | no limit | staff only |
| basic | 3 | Nvidia A100 GPU | 31 hours | for students |
| bigVoice | 2 | Nvidia A100 GPU | no limit | - |
| build | 3 | Nvidia A100 GPU | no limit | staff only |
The default queue for staff is doTrain (and basic for student) so it's not necessery to choose a queue, but it's possible to specify a different one.
Installed software and drivers
* NVIDIA A100 GPU drivers
* CuDA toolkit [version 11.7]
* Intel oneAPI Math Kernel Library
* Python 3.9.7
* pip 20.3.4
* ffmpeg + sox
If additional software is needed or different version, you can ask sysadmin (compute@ru.is) for assistance