![[2602.16719] GPU-Accelerated Algorithms for Graph Vector Search: Taxonomy, Empirical Study, and Research Directions](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.16719] GPU-Accelerated Algorithms for Graph Vector Search: Taxonomy, Empirical Study, and Research Directions
Summary
This paper presents a comprehensive survey of GPU-accelerated algorithms for graph vector search, detailing optimization strategies and empirical evaluations across various datasets.
Why It Matters
As data sizes increase, efficient retrieval methods like Approximate Nearest Neighbor Search (ANNS) are critical for machine learning applications. This study provides insights into optimizing graph-based approaches for modern GPU architectures, addressing performance bottlenecks and offering guidelines for future research.
Key Takeaways
- The study establishes a taxonomy of GPU optimization strategies for graph vector search.
- Distance computation is identified as a key bottleneck in performance.
- Data transfer between CPU and GPU significantly impacts real-world latency.
- A thorough evaluation of six algorithms on large-scale datasets is provided.
- Guidelines for designing scalable GPU-powered ANNS systems are outlined.
Computer Science > Databases arXiv:2602.16719 (cs) [Submitted on 10 Feb 2026] Title:GPU-Accelerated Algorithms for Graph Vector Search: Taxonomy, Empirical Study, and Research Directions Authors:Yaowen Liu, Xuejia Chen, Anxin Tian, Haoyang Li, Qinbin Li, Xin Zhang, Alexander Zhou, Chen Jason Zhang, Qing Li, Lei Chen View a PDF of the paper titled GPU-Accelerated Algorithms for Graph Vector Search: Taxonomy, Empirical Study, and Research Directions, by Yaowen Liu and 9 other authors View PDF HTML (experimental) Abstract:Approximate Nearest Neighbor Search (ANNS) underpins many large-scale data mining and machine learning applications, with efficient retrieval increasingly hinging on GPU acceleration as dataset sizes grow. Although graph-based approaches represent the state of the art in approximate nearest neighbor search, there is a lack of systematic understanding regarding their optimization for modern GPU architectures and their end-to-end effectiveness in practical scenarios. In this work, we present a comprehensive survey and experimental study of GPU-accelerated graph-based vector search algorithms. We establish a detailed taxonomy of GPU optimization strategies and clarify the mapping between algorithmic tasks and hardware execution units within GPUs. Through a thorough evaluation of six leading algorithms on eight large-scale benchmark datasets, we assess both graph index construction and query search performance. Our analysis reveals that distance computation rema...
