CUDA Out of Memory – Aurora ERA5 Deployment
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1k
- Forks
- 174
- PR merge metrics
- No merged PRs in 30d
Description
I'm deploying Aurora using the ERA5 example: https://microsoft.github.io/aurora/example_era5.html.
The deployment fails with the following error: RuntimeError: CUDA out of memory. Tried to allocate XX GiB. GPU 0 has a total capacity of XX GiB of which XX MiB is free...
I've attempted to reduce the rollout steps, but the issue persists. I understand that Azure VMs offer higher RAM options, but they are cost-prohibitive for my current needs.
So, what is the minimum GPU VRAM required to successfully run this Aurora example? Are there recommended configurations or optimizations to reduce GPU memory usage?
Any guidance would be appreciated.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the ERA5 example at https://microsoft.github.io/aurora/example_era5.html and reproduce the reported CUDA out-of-memory failure, recording the GPU configuration and rollout settings. Review the deployment guidance and issue details to determine whether the minimum VRAM and memory-saving configurations can be documented; done means the example has clear, verified resource requirements or optimization guidance.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100