SIVF让GPU上的向量索引实时更新，速度提升数十倍。
SIVF: GPU-Resident IVF Index for Streaming Vector Search
- 在GPU内存中直接修改索引，避免频繁跨设备传输数据。
- 删除操作延迟降低数量级，12卡集群达108万次/秒删除速率。
- 适合需要高速流式更新的推荐系统、搜索服务等场景。
GPU加速的倒排文件（IVF）索引是大规模向量搜索的工业标准，但依赖静态显存布局，难以实现实时更新。我们的基准分析表明，现有GPU IVF设计在流式场景下需昂贵的CPU-GPU数据传输来更新索引，导致系统延迟从毫秒级跃升至秒级。本文提出SIVF，一种原生运行于GPU的索引结构，通过一系列新数据结构与算法（如无冲突的滑块分配、非连续内存上的合并搜索），实现了高吞吐量的就地修改。SIVF已集成至开源向量搜索库Faiss。在多种向量数据集上对比基线的结果显示，其删除延迟相比当前最优方案降低数量级。分布式实验在12卡集群上验证了近乎完美的线性扩展性，达到407万向量/秒的总写入吞吐和108.5万向量/秒的删除吞吐。
原文摘要 · Abstract (English)
GPU-accelerated Inverted File (IVF) index is one of the industry standards for large-scale vector search but relies on static VRAM layouts that hinder real-time mutability. Our benchmark and analysis reveal that existing designs of GPU IVF necessitate expensive CPU-GPU data transfers for index updates, causing system latency to spike from milliseconds to seconds in streaming scenarios. We present SIVF, a GPU-native index that enables high-velocity, in-place mutation via a series of new data structures and algorithms, such as conflict-free slab allocation and coalesced search on non-contiguous memory. SIVF has been implemented and integrated into the open-source vector search library, Faiss. Evaluation against baselines with diverse vector datasets demonstrates that SIVF reduces deletion latency by orders of magnitude compared to the state-of-the-arts. Furthermore, distributed experiments on a 12-GPU cluster demonstrate that SIVF exhibits near perfect linear scalability, achieving an aggregate ingestion throughput of 4.07 million vectors/s and a deletion throughput of 108.5 million vectors/s.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。