Publication

gSampler: General and Efficient GPU-based Graph Sampling for Graph Learning

gSampler conducts a comprehensive study on 15 popular graph sampling algorithms to motivate a general and efficient GPU-based graph sampling framework. It employs the Extract-Compute-Select-Finalize (ECSF) model for single-layer graph sampling and provides matrix-centric APIs that are user-friendly and intuitive.

SOSP 2023 / October 2023

graph samplingGPUgraph learningGNN

PDF Code

Authors

Ping Gong, Renjie Liu, Zunyao Mao, Zhenkun Cai, Xiao Yan, Cheng Li, Minjie Wang, Zhuozhao Li

Abstract

gSampler uses the ECSF model to provide a general and efficient GPU-based graph sampling framework, unifying 15 popular sampling algorithms with matrix-centric APIs and achieving significant speedups over CPU-based approaches.