Medical Advanced Research Computing (MARC)
The Medical Advanced Research Computing (MARC) cluster is a high-performance/high-throughput computational resource to process research data classified as Level 4, based on the University of Calgary's Information Security Classification Standard. MARC is a general purpose, multiuser Linux computational cluster made up of multiple compute nodes to run jobs. Jobs can span more than one compute node if required.
The Medical Advanced Research Computing (MARC) service provides the features listed below, enabling research groups or individuals to perform data-intensive analyses on large datasets.
- Batch-oriented environment – Jobs are submitted and the job schedule runs the job when resources become available
- Jobs can span more than one compute node if required
- Provides a virtual environment where researchers can analyze their data and visualize results
- Packaged with Data Sharing service – Including Data Sharing and Collaboration (ShareFile)
- Provides compliance with ethics and privacy regulations for Level 4 data
- Login node for researchers to connect to
- 8 compute nodes with 2 V100 GPUs
- 40 cores
- 768 GB of RAM
- 1 large memory node
- 80 cores
- 3 TB of RAM
- 10 Gb network between all the compute nodes
- A central cluster file system
- Slurm batch job scheduler
- There is no cost associated with the use of this service
Before requesting access to MARC, users must have hands-on experience with:
- Linux command line
- Files and directories within a Linux environment
- File editor (ed, vi, vim, emacs)
- Job Scheduler (Slurm)
If you need more information, assistance, or training, please send an email to email@example.com.
Once you have access to a Secure Computing Data Storage (SCDS) group file share and have collected Level 4 data, you can request access to MARC. Allow two (2) business days for your account to be created.
If you have not yet requested your SCDS group file share be created or if you have not yet requested access to an existing SCDS group file share, find out how to do so on the Secure Computing Data Storage (SCDS) page.