Tenders Are Invited For Super Computer – Marum Dfg Cluster Of Excellence Project

Super Computer – Marum Dfg Cluster Of Excellence Project. The Subject Of The Award Is The Procurement Of It Hardware With A Service Component. Delivery Of Hardware To Build An Hpc Cluster, Ceph Storage Cluster And Gpu Nodes For Ai/Ml Applications. The Concept Of A Cross-Institutional Platform For Processing Large Amounts Of Data Makes It Possible For The First Time To Record, Process And Store Data In A System Environment. This Creates The Possibility Of Uniformly Processing Large Amounts Of Data In A High-Performance Computing Environment Using Classic Or Ai-Supported Algorithms, From Data Collection To Archiving. Modern Research Is Increasingly Characterized By The Generation And Analysis Of Large Amounts Of Data. The University Of Oldenburg (Uol), The University Of Bremen (Uhb) And The Alfred Wegener Institute (Awi) In Bremerhaven Have Strong Expertise In Data-Intensive Research Areas, Particularly In Marine Research (Uhb & Uol) And Geosciences (Abc). Currently, Fragmented Infrastructure And Decentralized Data Processing Limit The Efficiency And Scope Of This Research. Scientists Are Forced To Process Data Locally Or In Distributed Computing Centers, Leading To Redundancy, Incompatibilities And Limited Collaboration. The Existing Computing Power Is Often Insufficient For Complex Analyzes And Ai Applications. The Amount Of Data Processed Quickly Reaches Hundreds Of Terabytes Or Even Petabytes. The Processing Of Such Data Often Takes Place In A Multi-Stage Process: • Downloading Collected Research Data To A Local Processing Unit Or Transferring It To An Appropriately Equipped Computing Unit. • Processing The Data • Exchanging The Data With Other Researchers For Joint Work • Forwarding To A Publication System • Archiving The Data There Are Therefore Several Places Between Which The Data Must Be Moved. Collaboration In The Sense That Scientists Can Edit And Process The Data Together On One Platform Is Therefore Difficult To Achieve. The Present Concept Presents A Solution For A Consistent, Collaborative Platform For Processing Large Amounts Of Data That Can Be Used Interdisciplinary And Across Institutions.