Clusters
We have a number of compute clusters
- Pelle, reserved for Uppsala university
- Bianca, a part of NAISS-SENS, i.e. sensitive data
- UPPMAX Cloud, a part of Swedish Science Cloud
- Miarka, reserved for NGI's production environment
Technical summary of the UPPMAX clusters
Pelle | Bianca | |
|---|---|---|
Purpose | General-purpose | Sensitive |
CIA classification* | 222 | 322 |
# Intel CPU Nodes | 115 AMD | 288 |
# GPU Nodes | 4, 10x Nvidia L40s; 2, 2x Nvidia H100; 34, 1x Nvidia T4 | 10, 2x Nvidia A100 each |
Cores per node | 48 (96 threads) | 16 |
Memory per node | 768 GB | 128 GB |
Fat nodes | 2 TB & 3 TB | 256 & 512 GB |
Local disk (scratch) | 1.7 TB | 4 TB |
Login nodes | Yes | Yes (2 cores and 15 GB) |
"Home" storage | Gorilla | Cygnus |
"Project" Storage | Crex--> Gorilla | Cygnus |
* According to the Uppsala University information security guidelines.
Get an overview on the documentation pages as well.
Contact
- Use our support form! This is especially important for SENS projects and Bianca issues
- For Rackham/Pelle related questions you may still email support@uppmax.uu.se.
- For web matters, contact bjorn.claremar@uppmax.uu.se