iQAN: Fast and Accurate Vector Search with Efficient Intra-Query Parallelism on Multi-Core Architectures
Vector search has drawn a rapid increase of interest in the research community due to its application in novel AI applications. Maximizing its performance is essential for many tasks but remains preliminary understood. In this work, we investigate the root causes of the scalability bottleneck of using intra-query parallelism to speedup the state-of-the-art graph-based vector search systems on multi-core architectures. Our in-depth analysis reveals several scalability challenges from both system and algorithm perspectives. Based on the insights, we propose iQAN, a parallel search algorithm with a set of optimizations that boost convergence, avoid redundant computations, and mitigate synchronization overhead. Our evaluation results on a wide range of real-world datasets show that iQAN achieves up to $37.7\times$ and $76.6\times$ lower latency than state-of-the-art sequential baselines on datasets ranging from a million to a hundred million datasets. We also show that iQAN achieves outstanding scalability as the graph size or the accuracy target increases, allowing it to outperform the state-of-the-art baseline on two billion-scale datasets by up to $16.0\times$ with up to 64 cores.
Tue 28 FebDisplayed time zone: Eastern Time (US & Canada) change
15:40 - 16:40 | |||
15:40 20mTalk | iQAN: Fast and Accurate Vector Search with Efficient Intra-Query Parallelism on Multi-Core Architectures Main Conference Zhen Peng William & Mary, Minjia Zhang Microsoft Research, Kai Li Kent State University, Ruoming Jin Kent State University, Bin Ren College of William & Mary | ||
16:00 20mTalk | WISE: Predicting the Performance of Sparse Matrix Vector Multiplication with Machine Learning Main Conference Serif Yesil University of Illinois Urbana-Champaign, Azin Heidarshenas University of Illinois Urbana-Champaign, Adam Morrison Tel Aviv University, Josep Torrellas University of Illinois at Urbana-Champaign | ||
16:20 20mTalk | Efficient Direct Convolution Using Long SIMD Instructions Main Conference Alexandre Santana Barcelona Supercomputing Center, Adrià Armejach Sanosa Barcelona Supercomputing Center & Universitat Politècnica de Catalunya, Marc Casas Barcelona Supercomputing Center |