MELL combines adaptive token or KV-cache migration with an online scheduling algorithm that claims a 4/3 competitive ratio and reports 9-31% GPU savings in simulation.
Orca: A distributed serving system for Transformer-Based generative models,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.DC 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Mell: Memory-Efficient Large Language Model Serving via Multi-GPU KV Cache Management
MELL combines adaptive token or KV-cache migration with an online scheduling algorithm that claims a 4/3 competitive ratio and reports 9-31% GPU savings in simulation.