English

Learned Cost Model for Placement on Reconfigurable Dataflow Hardware

Distributed, Parallel, and Cluster Computing 2025-11-05 v1 Machine Learning Programming Languages

Abstract

Mapping a dataflow-graph of an ML model onto a reconfigurable system is difficult, as different mappings have different throughputs and consume resource constraints differently. To solve this, a model to evaluate the throughput of mappings is necessary as measuring throughput completely is expensive. Many use a hand-designed analytical model, relying on proxy features or intuition, introducing error. We provide a Learned Approach that predicts throughput 31%-52% more accurately over a variety of graphs. In addition, our approach shows no accuracy degradation after removing performance annotations. We show that using this approach results in 5.6% faster compiled graphs.

Keywords

Cite

@article{arxiv.2511.01872,
  title  = {Learned Cost Model for Placement on Reconfigurable Dataflow Hardware},
  author = {Etash Guha and Tianxiao Jiang and Andrew Deng and Jian Zhang and Muthu Annamalai},
  journal= {arXiv preprint arXiv:2511.01872},
  year   = {2025}
}

Comments

7 pages, 2 figures, 2 tables, DAC Conference style (2022)

R2 v1 2026-07-01T07:19:51.998Z