📋 Tender Overview
Analyzing And Exploiting Inexactness In Exascale Matrix Computations
Understanding The Limits Of Inexactness Could Lead To More Efficient Algorithms For Exascalematrix Computations Are Common In Many Fields. We Are Moving Towards The Age Of Exascale Computing – And Exascale-Size Matrices – Which Will Require Nearly Unfathomable Workloads. A Common Way To Reduce This Workload Is To Introduce Inexact Or Approximate Matrix Computations. But To Ensure That Inexactness From Multiple Sources Propagating Through Exascale Matrix Computations Does Not Confound The Conclusions We Draw From The Results, We Need A Systematic Approach For Analysing Multiple Sources Of Inexactness Together And Studying Their Interaction. Funded By The European Research Council, The Inexascale Project Will Rigorously Analyse The Combination Of Errors From Multiple Sources And Evaluate Their Effect On Numerical Behaviour And Solution Quality In Matrix Computations. This Could Lead To New Algorithms For Exascale Problems That Exploit Inexactness To Balance Performance And Accuracy.
Funded Under
European Research Council (Erc)
Total Cost
€ 1 496 085,00
Eu Contribution
€ 1 496 085,00
Coordinated By
Univerzita Karlova
Czechia
