Open access
Sep 2026
LEO: Enabling Efficient Communication-Computation Pipeline for GNN via Hierarchical Caching
The core of LEO lies in introducing hierarchical caching to construct a hybrid-grained software pipeline, enabling efficient computation-communication overlap both within and across GPU kernels.
Jia-Qi Si, De-Zun Dong
· ACM Transactions on Architec... · 0 citations