A data-driven framework predicts GNN inference latency for different accelerator dataflows and uses those predictions to schedule jobs, reporting large speedups in simulation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Data-Driven Approach to Dataflow-Aware Online Scheduling for Graph Neural Network Inference
A data-driven framework predicts GNN inference latency for different accelerator dataflows and uses those predictions to schedule jobs, reporting large speedups in simulation.