We have tested PageRank and Louvain on 1000+ GPUs. Many other algorithms (except for few with legacy implementations) should scale as well even though we haven’t tested every algorithm for 1000+ GPUs. If you need to run very large graph analytics and if your algorithm of interest does not scale well enough, pleae submit an issue in the cuGraph github page. For more academic reference, see Analyzing Multi-trillion Edge Graphs on Large GPU Clusters: A Case Study with PageRank | IEEE Conference Publication | IEEE Xplore
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Live AMA with cuGraph Engineering Team | 0 | 618 | April 12, 2023 | |
| About the AMA cuGraph: Graph analysis and GNN category | 3 | 759 | April 12, 2023 | |
| AMA with the cuGraph engineering team - April 12, 2023, 9am (PDT) | 7 | 2129 | April 12, 2023 | |
| Scaling large graphs | 2 | 528 | April 12, 2023 | |
| RAPIDS cuGraph: multi-GPU PageRank | 0 | 270 | August 21, 2022 | |
| Running Large-Scale Graph Analytics with Memgraph and NVIDIA cuGraph Algorithms | 2 | 368 | August 18, 2022 | |
| How does cuGraph support GNNs? | 3 | 525 | April 12, 2023 | |
| Beginner's Guide to GPU Accelerated Graph Analytics in Python | 1 | 512 | August 16, 2021 | |
| cuGraph: 그래프 분석 및 GNN 카테고리에 대해 무엇이든 물어보세요! | 0 | 590 | August 24, 2023 | |
| cuGraph pain points? | 2 | 536 | April 12, 2023 |