We have installed 80 general nodes to lonepeak. These are accessible without any allocation (= for free) and without preemption (= your job will not get deleted by higher priority jobs).
The catch is that the nodes are relatively old, comparable to Ember or existing Lonepeak 16 general nodes or slightly older. There is a mix of 8 core and 12 core nodes with 48-96 GB RAM, as detailed here:
There is also no InfiniBand network so these nodes would be ideal forindependent or loosely parallel jobs.
To access, simply ssh to lonepeak and modify your SLURM script as:
#SBATCH -A your_group_account
#SBATCH -p lonepeak
If you have any issues, please, let us know by sending email to firstname.lastname@example.org.