BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201805Z
LOCATION:241
DTSTART;TZID=America/Chicago:20251116T140000
DTEND;TZID=America/Chicago:20251116T142000
UID:submissions.supercomputing.org_SC25_sess209_ws_fgai116@linklings.com
SUMMARY:Exploring Distributed Vector Databases Performance on HPC Platform
 s: A Study with Qdrant
DESCRIPTION:Seth Ockerman (University of Wisconsin-Madison, Argonne Nation
 al Laboratory (ANL)); Amal Gueroudji (Argonne National Laboratory (ANL)); 
 Song Young Oh (University of Chicago); Robert Underwood and Nicholas Chia 
 (Argonne National Laboratory (ANL)); Kyle Chard (University of Chicago, Ar
 gonne National Laboratory (ANL)); Robert Ross (Argonne National Laboratory
  (ANL)); and Shivaram Venkataraman (University of Wisconsin-Madison)\n\nVe
 ctor databases have rapidly grown in popularity, enabling efficient simila
 rity search over data such as text, images, and video. They now play a cen
 tral role in modern AI workflows, aiding large language models by groundin
 g model outputs in external literature through retrieval-augmented generat
 ion. Despite their importance, little is known about the performance chara
 cteristics of vector databases in high-performance computing (HPC) systems
  that drive large-scale science. This work presents an empirical study of 
 distributed vector database performance on the  Polaris supercomputer in t
 he Argonne Leadership Computing Facility. We construct a realistic biologi
 cal-text workload from BV-BRC and generate embeddings from the peS2o corpu
 s using Qwen3-Embedding-4B. We select Qdrant to evaluate insertion, index 
 construction, and query latency with up to 32 workers. Informed by practic
 al lessons from our experience, this work takes a first step toward charac
 terizing vector database performance on HPC platforms to guide future rese
 arch and optimization.\n\nRecording: Livestreamed, Recorded\n\nRegistratio
 n Category: Technical Program Reg Pass, Workshop Reg Pass\n\nSession Chair
 s: Charlie Catlett (Argonne National Laboratory (ANL), University of Chica
 go); Javier Aula-Blasco (Barcelona Supercomputing Center (BSC)); Valerie T
 aylor (Argonne National Laboratory (ANL), University of Chicago); Laura Mo
 rselli (CINECA); and Rio Yokota (Institute of Science Tokyo)\n\n
END:VEVENT
END:VCALENDAR