Home

The Office of Research Computing administers a 1500+ CPU core batch-processing cluster called ARGO and a new cluster called HOPPER (currently with 3680 cores) available to all GMU faculty and faculty-sponsored students.
The ARGO and HOPPER clusters are both based on the condo model. However there are differences in how this "condo" model is implemented:
- Contributors buying nodes for the ARGO clusters are provided higher priority access to the whole cluster and temporary exclusive access to the funded nodes by arrangement.
- Contributors buying nodes for the HOPPER clusters are provided preemptive access to funded nodes.
This wiki provides login information and a number of HOWTO's as a guide to ORC users.
Articles
- About ARGO
- About HOPPER
- Important Differences between ARGO and HOPPER
- Getting an ORC Cluster Account
- Requesting Help from the Support Staff
- Logging into ARGO
- Logging into HOPPER
- Uploading Data
- Avoiding Quota Issues on Shared projects and group Directories
- Getting Started with SLURM
- Environment Modules
- Installing User Packages in ARGO
- Python Virtual Environment
- Intel Python
- How to Run MATLAB on ARGO
- How to Run R
- How to Manage R Packages
- How to Run Parallel Jobs on ARGO
- How to Write Parallel Python Code
- How to Run NAMD Jobs on ARGO
- How to Run Macaulay2 on ARGO
- Running CUDA on ARGO
- How to Run Tensorflow/Keras Jobs on ARGO
- How to Run Caffe Jobs on ARGO
- Creating Checkpoints (DMTCP)
- Using Globus
- ARGO Best Practices Q&A
- ARGO Frequently Asked Questions
- HOPPER DGX_A100_User_Guide
- Open OnDemand on HOPPER