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:20260202T201248Z
LOCATION:Second Floor Atrium
DTSTART;TZID=America/Chicago:20251120T080000
DTEND;TZID=America/Chicago:20251120T170000
UID:submissions.supercomputing.org_SC25_sess533_post268@linklings.com
SUMMARY:Applying Lossy Compression Techniques to GNN Training
DESCRIPTION:Milan Shah, Reece Neff, and Michela Becchi (North Carolina Sta
 te University)\n\nGraph neural networks (GNNs) are a state-of-the-art mach
 ine learning model for processing graph-structured data. The growing compl
 exity of GNNs and size of real-world graphs have increased the memory requ
 irements of GNN training and popular training platforms, like GPU, have me
 mory capacity on the scale of tens of GB. \n\nIn this work, we study scien
 tific floating-point lossy compressors applied to GNN training memory redu
 ction. We develop a framework for GNN activation lossy compression, analyz
 e lossy compression and other data reduction techniques, and explore metho
 ds to leverage GNN data features to improve compression. This work is ongo
 ing and will encompass more compression optimizations in the future.\n\nTh
 e poster session will provide an overview of GNN training and opportunitie
 s for compression, followed by an analysis of cuSZp, a scientific float-po
 int lossy compressor, GNN performance against quantization and reduced pre
 cision, and lastly, preliminary exploration of leveraging GNN attributes f
 or compression with top-k methods.\n\nTag: Research & ACM SRC Posters\n\nR
 egistration Category: Technical Program Reg Pass\n\nSession Chairs: Kento 
 Sato (RIKEN Center for Computational Science (R-CCS)); Chris Schlipalius (
 Pawsey Supercomputing Research Centre; Commonwealth Scientific and Industr
 ial Research Organisation (CSIRO), Australia); and Anja Gerbes (Georg-Augu
 st-Universität Göttingen)\n\n
END:VEVENT
END:VCALENDAR