ReAct-style LLM schedulers can balance multiple HPC scheduling objectives on 10-100 job workloads, though cloud API latency makes them unsuitable for real-time deployment.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Evaluating the Efficacy of LLM-Based Reasoning for Multiobjective HPC Job Scheduling
ReAct-style LLM schedulers can balance multiple HPC scheduling objectives on 10-100 job workloads, though cloud API latency makes them unsuitable for real-time deployment.