Skip to content
UCT eResearch Documentation
GPUs & Accelerators
Initializing search
UCT eResearch Documentation
HIGH-PERFORMANCE COMPUTING (HPC)
Start Here
Start Here
Quick Start
Access & Accounts
Logging In
Running Your First Job
Basic Linux You’ll Use
Do Work (Run Jobs)
Do Work (Run Jobs)
Scheduler Basics
Submit Batch Jobs
Interactive Jobs
Arrays & Workflows
GPUs & Accelerators
Monitoring & Efficiency
Software & Environments
Software & Environments
Modules (Basics)
Python on HPC
R on HPC
Compilers & Math Libs
Containers (Apptainer)
Software Catalog
Data & Storage
Data & Storage
Filesystems & Quotas
Move Data (scp/rsync/Globus)
Manage Your Data
Archive & Retention
Databases & Datasets
Interactive & Portals
Interactive & Portals
Open OnDemand
Dashboards
Account/Allocations Portal
Systems & Hardware
Systems & Hardware
System Overview
Login Nodes
Compute Nodes
Partitions/QoS
Troubleshooting & FAQs
Troubleshooting & FAQs
FAQs
Common Job Failures
Debugging Jobs
Connection Problems
Who to Contact
Performance & Optimization
Performance & Optimization
Profiling & Tracing Tools
MPI/OpenMP/Hybrid
GPU Performance
I/O Best Practices
Policies & Governance
Policies & Governance
User Policy & AUP
Security
Allocations
Maintenance
Citing the Facility
Investment Options
Training & How-tos
Training & How-tos
Courses
Short Videos
Guided Tutorials
Tips & Tricks
Reference
Reference
Slurm Directives Cheat Sheet
Linux Commands Quick Reference
Glossary
What's New
GPUs & Accelerators
¶
Choose GPU partitions and run CUDA/ROCm jobs with working examples.