BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143141Z
LOCATION:B306
DTSTART;TZID=America/New_York:20241118T140000
DTEND;TZID=America/New_York:20241118T142300
UID:submissions.supercomputing.org_SC24_sess751_ws_p3hpc102@linklings.com
SUMMARY:Performance Modeling and Analysis of a de Bruijn Graph Based Local
  Assembly Kernel on Multiple Vendor GPUs
DESCRIPTION:LeAnn Lindsey (University of Utah) and Nan Ding, Jack DeSlippe
 , and Muaaz Awan (Lawrence Berkeley National Laboratory (LBNL))\n\nBioinfo
 rmatics workloads differ significantly from traditional scientific computi
 ng and AI workloads because they consist primarily of integer-only operati
 ons and string comparisons rather than floating-point operations. The unde
 rlying algorithms usually have low arithmetic intensity, irregular memory 
 access patterns, and non-deterministic workloads. Local Assembly is an ess
 ential step in large-scale genome assembly software and is typically imple
 mented using de Bruijn graphs. This paper examines the performance, portab
 ility, and productivity of a local assembly GPU kernel from a metagenome a
 ssembly pipeline implemented using hash table data structures on NVIDIA, A
 MD, and Intel GPUs. We focus on the challenges of achieving portability wh
 ile maintaining performance for a complex bioinformatics GPU kernel that r
 elies on hardware-specific optimizations. In this paper, we evaluate the l
 ocal assembly kernel's performance and portability across different GPU ar
 chitectures, identify performance bottlenecks, and propose modifications i
 n existing tools and methods for performance modeling and analysis of inte
 ger-heavy bioinformatics application kernels.\n\nTag: Performance Optimiza
 tion, Programming Frameworks and System Software\n\nRegistration Category:
  Workshop Reg Pass\n\nSession Chairs: CJ Newburn (NVIDIA Corporation), Sco
 tt J. Parker (Argonne National Laboratory (ANL)), John Pennycook (Intel Co
 rporation), and Kenneth Weiss (Lawrence Livermore National Laboratory (LLN
 L))\n\n
END:VEVENT
END:VCALENDAR