1 Hit in 1.2 sec

FairGV: Fair and Fast GPU Virtualization

Cheol-Ho Hong, Ivor Spence, Dimitrios S. Nikolopoulos
2017 IEEE Transactions on Parallel and Distributed Systems  
This paper presents FairGV, a radically redesigned GPU virtualization system that achieves system-wide weighted fair sharing and strong performance isolation in mixed workloads that use GPUs with variable  ...  To achieve its objectives, FairGV introduces a trap-less GPU processing architecture, a new fair queuing method integrated with work-conserving and GPU-centric coscheduling polices, and a collaborative  ...  For realizing fair and efficient GPU virtualization, we introduce FairGV, a new system that combines a highly optimized GPU virtualization framework with a novel fair-share scheduler.  ... 
doi:10.1109/tpds.2017.2717908 fatcat:d3ik4axiqjgjjmwfhvd6uc36ku