Poster Type: Research Posters
Author: Seth Ockerman (University of Wisconsin-Madison), Song Young Oh (University of Chicago), Amal Gueroudji (Argonne National Laboratory), Rochana Chaturvedi (Ashoka University), Philip Carns (Argonne National Laboratory), Nicholas Chia (Argonne National Laboratory), Matthieu Dorier (Argonne National Laboratory), Robert Latham (Argonne National Laboratory), Tanwi Mallick (Argonne National Laboratory), Swann Perarnau (Argonne National Laboratory), Robert Underwood (Argonne National Laboratory), Kyle Chard (University of Chicago, Argonne National Laboratory), Ian Foster (Argonne National Laboratory, University of Chicago), Robert Ross (Argonne National Laboratory), Shivaram Venkataraman (University of Wisconsin-Madison)
Abstract: Vector databases have been designed and optimized for cloud environments; however, emerging scientific AI workloads (e.g., molecular search, meteorological trajectory detection, and literature-driven hypothesis generation) demand efficient, scalable execution on HPC systems. We present a large-scale evaluation of three state-of-the-art vector databases — Qdrant, Milvus, and Weaviate — on two production supercomputers, scaling to 256 distributed workers across 64 compute nodes. We evaluate VDBs using Yandex-T2I, a multimodal embedding dataset, and Pes2o-VE, a novel real-world scientific dataset. Our results reveal that insertion is bottlenecked by per-batch consistency requirements, additional cores can reduce query throughput by up to 30.67%, and scaling from 16 to 256 workers yields a 2.67x improvement with 16 nodes before QPS begins to decline. This scaling paradox exposes the fundamental mismatch between cloud-oriented designs and HPC systems, highlighting the need for new, HPC-aware vector database designs.
Best Poster Finalist (BP): no
Poster: PDF
Poster Summary: PDF