GPart: A GNN-Enabled Multilevel Graph Partitioner

This paper introduces GPart, a scalable multilevel framework for graph partitioning that integrates GNN embeddings with efficient coarsening and refinement techniques. On the Titan23 benchmarks, GPart achieves a cut size reduction of 34.13% to 42.92% over METIS and improves cut size by 9.30% on sele...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	2025 62nd ACM/IEEE Design Automation Conference (DAC) s. 1 - 7
Hlavní autoři:	Chen, Magi, Wang, Ting-Chi
Médium:	Konferenční příspěvek
Jazyk:	angličtina
Vydáno:	IEEE 22.06.2025
Témata:	Benchmark testing Design automation graph neural network Graph neural networks graph partitioning Memory architecture Memory management Scalability Training
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	This paper introduces GPart, a scalable multilevel framework for graph partitioning that integrates GNN embeddings with efficient coarsening and refinement techniques. On the Titan23 benchmarks, GPart achieves a cut size reduction of 34.13% to 42.92% over METIS and improves cut size by 9.30% on selected DIMACS benchmarks compared to G-kway. Furthermore, experiments on the Titan23 benchmarks show that GPart reduces normalized memory usage by 24.6x compared to GAP and 12.4x compared to GenPart. Unlike existing GNN-based methods, which require large hidden layers and substantial memory, GPart's multilevel architecture reduces hidden layer sizes, significantly optimizing memory efficiency.
DOI:	10.1109/DAC63849.2025.11132100