The use of general-purpose GPUs nodes to off-load image processing from the main CPU are now common practice in computational research. Arrays of general-purpose GPU nodes are no longer considered as niche infrastructure but an expected component of computational services for research-based applications.
The key ECDF services, Eddie (large scale computation) and Eleanor (private cloud), have incorporated GPU subsystems in their infrastructure for the last 5 years. These nodes have now reached end-of-life and need to be replaced. In addition, modern research computation requires arrays of GPU nodes to support high-demand applications in areas such as artificial intelligence (AI) and machine learning (ML).
This project will deliver the necessary upgrade to the GPU environment and facilitate the expansion of this service. Additional project documentation can be found in the project SharePoint area.
Current project status
|Report Date||RAG||Budget||Effort Completed||Effort to complete|
|May 2023||BLUE||35.0 days||73.0 days||0.0|