# Distributed Graph Neural Network Inference With Just-In-Time Compilation For Industry-Scale Graphs

arXiv: https://arxiv.org/abs/2503.06208 · HTML: https://mlsanity.com/paper/2503.06208/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 5 references that are in the map.
- First posted on arXiv 8 Mar 2025.

## Builds on (most cited first, 4 of 4)

- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [Fast Graph Representation Learning with PyTorch Geometric](https://mlsanity.com/paper/1903.02428/) (2019)
- [GraphTheta: A Distributed Graph Neural Network Learning System With Flexible Training Strategy](https://mlsanity.com/paper/2104.10569/) (2021)
- [DFOGraph: An I/O- and Communication-Efficient System for Distributed Fully-out-of-Core Graph Processing](https://mlsanity.com/paper/2101.06911/) (2021)

## Built on by (most cited first, 1 of 1)

- [GraphFAS: A Distributed System for Automated Graph Feature Generation and Selection in Industrial Transaction Networks](https://mlsanity.com/paper/2609.08970/) (2026)

## Abstract (arXiv)

Graph neural networks (GNNs) have delivered remarkable results in various fields. However, the rapid increase in the scale of graph data has introduced significant performance bottlenecks for GNN inference. Both computational complexity and memory usage have risen dramatically, with memory becoming a critical limitation. Although graph sampling-based subgraph learning methods can help mitigate computational and memory demands, they come with drawbacks such as information loss and high redundant computation among subgraphs. This paper introduces an innovative processing paradgim for distributed graph learning that abstracts GNNs with a new set of programming interfaces and leverages Just-In-Time (JIT) compilation technology to its full potential. This paradigm enables GNNs to highly exploit the computational resources of distributed clusters by eliminating the drawbacks of subgraph learning methods, leading to a more efficient inference process. Our experimental results demonstrate that on industry-scale graphs of up to 500 million nodes and 22.4 billion edges, our method can produce a performance boost of up to 27.4 times.