LaissezCloud allows ongoing bidding for cloud resources, cutting performance loss from contention by 8-23% over standard on-demand and spot methods while scaling to 10,000 nodes.
Orion: Interference- aware, fine-grained gpu sharing for ml applications
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
LaissezCloud: Continuous Resource Renegotiation for the Public Cloud
LaissezCloud allows ongoing bidding for cloud resources, cutting performance loss from contention by 8-23% over standard on-demand and spot methods while scaling to 10,000 nodes.