Welcome to the Office of Advanced Research Computing at Rutgers!
OARC is a university-wide initiative that aims to develop and implement a strategic vision for centralizing the advanced research computing and data cyberinfrastructure (ACI) ecosystem at Rutgers. OARC has the goal of providing Rutgers researchers with essential computing and data handling capabilities, and students with necessary exposure and training, through centralized resources, services and training.
These pages are a collection of resources to help you to utilize the cluster more effectively. Even if you are a very experienced Linux user, you will want to read cluster user guide as it has slurm tips and examples.
WARNING - READ!
- Do not run large computational jobs on the login node. Use slurm to allocate resources on the compute node. Failure to respect the golden rule can get your account suspended.
- Number of jobs to submit should not exceed 5,000 at a time - queue is currently 10,000 and slurm will go to sleep if that is exceeded.
- Right-size your jobs: jobs may run sooner if the resources requested are smaller and your future priority is degraded if you have asked for (even if not used) a lot of resources lately.
-noption in sbatch only if you know what you are doing. Multithreaded but not multi-node jobs should use
- TODO: add more complete list
- If you understand material in lmod cheatsheet and slurm cheatsheet you are good to go.
- For users familiar with Linux but new to
slurm, check out environment modules and intro to slurm.
- For users not familiar with Linux, please familiarize yourself with Linux through tutorials and cheatsheets, or use OnDemand until you feel more comfortable with Linux. If you wish to learn about HPC in detail, check out HPC Carpentry course for beginners.