Difference between revisions of "CCU:Roadmap"
Jump to navigation
Jump to search
m |
|||
| Line 18: | Line 18: | ||
:- set up Kubernetes as a cluster scheduler | :- set up Kubernetes as a cluster scheduler | ||
:- link to GPU nodes and make them available | :- link to GPU nodes and make them available | ||
| + | |||
* Compute servers for machine learning | * Compute servers for machine learning | ||
| − | :- GPU node nVidia DGX-2 (16 x V100) | + | :- GPU node nVidia DGX-2 (16 x V100) will be installed 27.5., available with full capacity on 11.6. |
| − | :- GPU node with 4x Titan RTX has | + | :- GPU node with 4x Titan RTX has arrived (reserved for SFB TRR 161) and is being set up |
:- Existing multi-GPU systems might be integrated in the cluster in the future. | :- Existing multi-GPU systems might be integrated in the cluster in the future. | ||
Revision as of 13:30, 17 May 2019
This is a general roadmap for the CCU, rough timelines will be added later on.
Wikimedia site
- Finish help pages for projects
- Finish help pages for computational resources (once those are available and all systems are in place)
Computational resources
- CCU access server has been ordered, once this is here:
- - set up new user authentication systems
- - move Wiki to new server
- - set up git and container repositories
- - set up data storage
- - set up Kubernetes as a cluster scheduler
- - link to GPU nodes and make them available
- Compute servers for machine learning
- - GPU node nVidia DGX-2 (16 x V100) will be installed 27.5., available with full capacity on 11.6.
- - GPU node with 4x Titan RTX has arrived (reserved for SFB TRR 161) and is being set up
- - Existing multi-GPU systems might be integrated in the cluster in the future.
