---
格式版本: 2
标题: "FROG: Efficient Range-Filtering Approximate Nearest Neighbor Search on GPUs"
原文链接: "https://arxiv.org/abs/2608.16491"
发布日期: "2026-08-17"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Mon, 17 Aug 2026 12:29:11 UTC (720 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:47:49+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:43:02+08:00"
入库时间: "2026-08-20T07:47:49.802Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=throughput&searchtype=all"
匹配关键词:
  - "throughput"
  - "GPU"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 25
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "内容为GPU数据库查询算法论文，仅涉及GPU与throughput，与超节点、AI Rack、机柜级AI基础设施及相关部件、供电、散热、互连、量产落地等主题完全无关，判定无关。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T15:54:19+08:00"
AI主题相关性: 0
AI来源权威性: 10
AI新颖性: 5
AI技术细节: 5
AI商业部署信号: 0
AI完整性: 5
AI摘要: "FROG提出一种面向GPU的范围过滤近似最近邻搜索（RFANNS）索引，采用全局感知的顶点中心设计，提升查询吞吐与构建效率。实验显示，相比44核CPU基线和最强GPU基线，其混合选择性查询吞吐分别提升14.7—37.7倍和4.5—7.6倍。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:24:39.253Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.16491"
---

## Computer Science > Databases

## Title:FROG: Efficient Range-Filtering Approximate Nearest Neighbor Search on GPUs

Authors:[Xiaokun Cui](https://arxiv.org/search/cs?searchtype=author&query=Cui,+X), [Pengbo Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+P), [Jiadong Xie](https://arxiv.org/search/cs?searchtype=author&query=Xie,+J), [Yingfan Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+Y), [Hui Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+H), [Jeffrey Xu Yu](https://arxiv.org/search/cs?searchtype=author&query=Yu,+J+X), [Jiangtao Cui](https://arxiv.org/search/cs?searchtype=author&query=Cui,+J)

[View PDF](https://arxiv.org/pdf/2608.16491) [HTML (experimental)](https://arxiv.org/html/2608.16491v1)

> Abstract:Range-filtering approximate nearest neighbor search (RFANNS) is a fundamental operation in modern vector databases. Given a query vector $q$ and a numerical range predicate, RFANNS returns the $k$-approximate nearest neighbors ($k$-ANN) of the query $q$ among the objects whose attributes satisfy the range predicate. However, existing RFANNS methods are not well suited to high-throughput GPU execution. CPU indexes offer limited parallel scalability, generic GPU filtering is highly selectivity-dependent, and GPU indexes built from locally optimized subgraphs can incur long search trajectories and redundant distance computations. To address these limitations, we present FROG, a GPU-oriented RFANNS index that replaces multiple locally optimal substructure building with a globally aware, vertex-centric design. It organizes diverse expansion neighbor candidates for each vertex in a GPU-friendly structure and rapidly identifies the expansion neighbors used for computation at query time. Moreover, GPU-oriented algorithms and implementations are developed for both index construction and query processing. Experiments on six datasets show that FROG improves mixed-selectivity query throughput by 14.7--37.7 $\times$ over 44-core CPU baselines and 4.5--7.6 $\times$ over the strongest GPU baseline. It also accelerates index construction by 2.4--14.8 $\times$ over the GPU baseline.

| Subjects: | Databases (cs.DB); Information Retrieval (cs.IR) |
| --- | --- |
| Cite as: | [arXiv:2608.16491](https://arxiv.org/abs/2608.16491) \[cs.DB\] |
|  | (or [arXiv:2608.16491v1](https://arxiv.org/abs/2608.16491v1) \[cs.DB\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.16491](https://doi.org/10.48550/arXiv.2608.16491) |

## Submission history

From: Yingfan Liu \[[view email](https://arxiv.org/show-email/0fe61518/2608.16491)\]  
**\[v1\]** Mon, 17 Aug 2026 12:29:11 UTC (720 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.16491) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))