SVFusion: A CPU-GPU Co-Processing Architecture for Large-Scale Real-Time Vector Search
Summary: SVFusion co-designs a hierarchical GPU–CPU–disk ANN index for high-throughput search with online updates, using workload-aware GPU caching, CUDA streams, adaptive resource management, and concurrency control. It delivers up to 20.9× higher throughput and 50.7× lower latency while preserving recall. (summarized by gpt-5.6-luna on Jul 24 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yuchen Peng (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
- 2. Dingyu Yang (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
- 3. Zhongle Xie (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
- 4. Ji Sun (Huawei)
- 5. Lidan Shou (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
- 6. Ke Chen (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
- 7. Gang Chen (Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security; Zhejiang University)
BibTeX Citation
@article{peng_vldb26,
title = {{SVFusion: A CPU-GPU Co-Processing Architecture for Large-Scale Real-Time Vector Search}},
author = {Peng, Yuchen and Yang, Dingyu and Xie, Zhongle and Sun, Ji and Shou, Lidan and Chen, Ke and Chen, Gang},
journal = {PVLDB},
series = {{VLDB} '26},
volume = {19},
number = {5},
pages = {1074--1087},
doi = {10.14778/3796195.3796216},
url = {https://doi.org/10.14778/3796195.3796216},
year = {2026}
}
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 16 of 16 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 7,474 | GTS: GPU-based Tree Index for Fast Similarity Search | 2024 | SIGMOD |
| 2 | 10,036 | VStream: A Distributed Streaming Vector Search System | 2025 | VLDB |
| 3 | 9,778 | Turbocharging Vector Databases using Modern SSDs | 2025 | VLDB |
| 4 | 10,456 | FlashANNS: GPU-Driven Asynchronous I/O Pipelining for Eliminating Storage-Compute Bottlenecks in Billion-Scale Similarity Search | 2026 | SIGMOD |
| 5 | 7,166 | Accelerating Graph Indexing for ANNS on Modern CPUs | 2025 | SIGMOD |
| 6 | 13,290 | VecFlow-Chamfer: A GPU-based Data Management System for High-Performance Multi-Vector Search on Superchips | 2026 | SIGMOD |
| 7 | 10,449 | Efficient Vector Index Merging in Vector Databases | 2026 | SIGMOD |
| 8 | 10,227 | Efficient Index Layout and Search Strategy for Large-scale High-dimensional Vector Similarity Search | 2026 | SIGMOD |
| 9 | 10,333 | VecFlow: A High-Performance Vector Data Management System for Filtered-Search on GPUs | 2026 | SIGMOD |
| 10 | 10,377 | High-Throughput, Cost-Effective Billion-Scale Vector Search with a Single GPU | 2026 | SIGMOD |