The paper develops fluid-guided online scheduling algorithms (WAIT and Nested WAIT) for LLM inference that handle endogenous KV-cache memory growth and improve stability and latency over baselines in simulations.
Negative drift ensures queue stability, preventing unbounded growth that would lead to eviction
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Optimizing LLM Inference: Fluid-Guided Online Scheduling with Memory Constraints
The paper develops fluid-guided online scheduling algorithms (WAIT and Nested WAIT) for LLM inference that handle endogenous KV-cache memory growth and improve stability and latency over baselines in simulations.