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:20260202T201808Z
LOCATION:261-262-265-266
DTSTART;TZID=America/Chicago:20251120T143700
DTEND;TZID=America/Chicago:20251120T150000
UID:submissions.supercomputing.org_SC25_sess162_pap374@linklings.com
SUMMARY:lsCOMP: Efficient Light Source Compression
DESCRIPTION:Yafan Huang (University of Iowa); Sheng Di, Robert Underwood, 
 Peco Myint, and Miaoqi Chu (Argonne National Laboratory (ANL)); Guanpeng L
 i (University of Florida); and Nicholas Schwarz and Franck Cappello (Argon
 ne National Laboratory (ANL))\n\nLight source facilities, which generate X
 -rays for probing microstructures and dynamic processes, produce intense d
 ata streams, reaching up to 250 GB/s and projected to exceed 1 TB/s by the
  end of this decade. Managing such massive data poses critical challenges 
 due to limited local processing capacity and bandwidth constraints when of
 floading data to HPC systems. To address these challenges, we propose lsCO
 MP, a GPU compressor that operates within a single kernel. lsCOMP supports
  both lossless and configurable lossy compression, ensuring high compressi
 on ratios and preserved data quality across diverse light source applicati
 ons. On one NVIDIA A100 GPU, lsCOMP achieves compression throughputs of 38
 0.89 to 509.21 GB/s in lossless mode, delivering up to 20 times higher per
 formance than industry-leading GPU compressors while achieving superior co
 mpression ratios. In lossy modes, lsCOMP further improves throughput and r
 atios significantly. Additionally, lsCOMP demonstrates versatile performan
 ce across various integer datasets and supports TB/s-level random access t
 hroughput.\n\nTag: Algorithms, Applications, State of the Practice\n\nReco
 rding: Livestreamed, Recorded\n\nRegistration Category: Technical Program 
 Reg Pass\n\nSession Chair: Shaikh Arifuzzaman (University of Nevada, Las V
 egas)\n\n
END:VEVENT
END:VCALENDAR