PolarFly: A Cost-Effective and Flexible Low-Diameter Topology
Kartik Lakhotia, Maciej Besta, Laura Monroe, Kelly Isham, Patrick Iff, Torsten Hoefler, Fabrizio Petrini
摘要
In this paper we present PolarFly, a diameter-2 network topology based on the Erdős-Rényi family of polarity graphs from finite geometry. This is a highly scalable lowdiameter topology that asymptotically reaches the Moore bound on the number of nodes for a given network degree and diameter.
PolarFly achieves high Moore bound efficiency even for the moderate radixes commonly seen in current and near-future routers, reaching more than 96% of the theoretical peak. It also offers more feasible router degrees than the state-of-the-art solutions, greatly adding to the selection of scalable diameter-2 networks. PolarFly enjoys many other topological properties highly relevant in practice, such as a modular design and expandability that allow incremental growth in network size without rewiring the whole network. Our evaluation shows that PolarFly outperforms competitive networks in terms of scalability, cost and performance for various traffic patterns.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了最后一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper6
- Octopus: Enhancing CXL Memory Pods via Sparse TopologyYuhong Zhong, Fiodar Kazhamiaka, Pantea Zardoshti, Shuwei Teng 等NSDI 2026 · 被引用 15 次
- A High-Performance Design, Implementation, Deployment, and Evaluation of The Slim Fly NetworkNils Blach, Maciej Besta, Daniele De Sensi, Jens Domke 等NSDI 2024 · 被引用 13 次
- Switch-Less Dragonfly on Wafers: A Scalable Interconnection Architecture based on Wafer-Scale IntegrationYinxiao Feng, Kaisheng MaSC 2024 · 被引用 10 次
- NetSparse: In-Network Acceleration of Distributed Sparse KernelsGerasimos Gerogiannis, Dimitrios Merkouriadis, Charles Block, Annus Zulfiqar 等MICRO 2025 · 被引用 1 次
- Starfish: A Topology-Routing Co-Design for Small-Scale Data CentersAnchengcheng Zhou, Vipul Harsh, Sangeetha Abdu Jyothi, Maria ApostolakiNSDI 2026
它引用的顶会 Paper3
- ZeRO-infinity: breaking the GPU memory wall for extreme scale deep learningSamyam Rajbhandari, Olatunji Ruwase, Jeff Rasley, Shaden Smith 等SC 2021 · 被引用 254 次
- Clio: a hardware-software co-designed disaggregated memory systemZhiyuan Guo, Yizhou Shan, Xuhao Luo, Yutong Huang 等ASPLOS 2022 · 被引用 110 次
- FatPaths: routing in supercomputers and data centers when shortest paths fall shortMaciej Besta, Marcel Schneider, Marek Konieczny, Karolina Cynk 等SC 2020 · 被引用 25 次
相关 Paper
- Zettafly: A Network Topology with Flexible Non-blocking Regions for Large-scale AI and HPC SystemsDezun Dong, Ziyu Wang, Fei LeiISCA 2025 · 被引用 7 次
- Architecture and performance studies of 3D-Hyper-FleX-LION for reconfigurable all-to-all HPC networksGengchen Liu, Roberto Proietti, Marjan Fariborz, Pouya Fotouhi 等SC 2020 · 被引用 21 次
- TAGO: rethinking routing design in high performance reconfigurable networksMin Yee Teh, Yu-Han Hung, George Michelogiannakis, Shijia Yan 等SC 2020 · 被引用 5 次
- An Evaluation of the Effect of Network Cost Optimization for Leadership Class SupercomputersAwais Khan, John R. Lange, Nick Hagerty, Edwin F. Posada 等SC 2024 · 被引用 4 次
- Balanced Sparse Tree: A Scalable Network Topology for Large Language ModelsShaoteng Liu, Dejun Kong, Huitian Wang, Hongji Dong 等SIGCOMM 2026
