动态图静态与增量扩展 DF-P PageRank 的高效 GPU 实现
Efficient GPU Implementation of Static and Incrementally Expanding DF-P PageRank for Dynamic Graphs
AI总结:
本文提出了高效的静态和动态图 DF-P PageRank 的 GPU 实现,通过基于拉取的无原子计算和顶点度数划分优化性能,在 NVIDIA A100 上显著超越现有方法,并在动态图上实现 2.1-3.1 倍加速。
摘要:
PageRank 是一种广泛使用的中心性度量,它通过考虑连接及其重要性来对图中的顶点进行“排名”。在本报告中,我们首先介绍了 Static PageRank 最高效的 GPU 实现之一,该方法从头重新计算 PageRank 分数。它采用同步的基于拉取的、无原子操作的 PageRank 计算方式,并将低入度和高入度顶点进行划分,由两个独立的内核处理。接下来,我们提出了增量扩展(和收缩)Dynamic Frontier with Pruning (DF-P) PageRank 的 GPU 实现,该方法仅处理可能改变排名的顶点子集。它基于 Static PageRank,并在低出度和高出度顶点之间使用额外的划分,通过两个额外的内核实现受影响顶点集合的增量扩展。在配备 NVIDIA A100 GPU 的服务器上,我们的 Static PageRank 分别比 Hornet 和 Gunrock 的 PageRank 实现快 31 倍和 5.9 倍。在此基础上,DF-P PageRank 在真实世界动态图上比 Static PageRank 快 2.1 倍,在带有随机批量更新的大型静态图上快 3.1 倍。
英文摘要:
PageRank is a widely used centrality measure that "ranks" vertices in a graph by considering the connections and their importance. In this report, we first introduce one of the most efficient GPU implementations of Static PageRank, which recomputes PageRank scores from scratch. It uses a synchronous pull-based atomics-free PageRank computation, with the low and high in-degree vertices being partitioned and processed by two separate kernels. Next, we present our GPU implementation of incrementally expanding (and contracting) Dynamic Frontier with Pruning (DF-P) PageRank, which processes only a subset of vertices likely to change ranks. It is based on Static PageRank, and uses an additional partitioning between low and high out-degree vertices for incremental expansion of the set of affected vertices with two additional kernels. On a server with an NVIDIA A100 GPU, our Static PageRank outperforms Hornet and Gunrock's PageRank implementations by 31x and 5.9x respectively. On top of the above, DF-P PageRank outperforms Static PageRank by 2.1x on real-world dynamic graphs, and by 3.1x on large static graphs with random batch updates.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。