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:B206
DTSTART;TZID=America/New_York:20241121T140000
DTEND;TZID=America/New_York:20241121T143000
UID:submissions.supercomputing.org_SC24_sess702_exforum143@linklings.com
SUMMARY:Towards Performance Portability on Cloud Clusters: Lessons Learned
  from Benchmarking MPI Applications
DESCRIPTION:Stefan Gary (Parallel Works)\n\nCloud-based HPC empowers pract
 itioners with more flexibility than traditional on-premise systems; on-dem
 and, customizable compute resources can be placed close to big data, thus 
 minimizing I/O bottlenecks and allowing users to explore cutting-edge hard
 ware. However, deploying HPC applications to different environments while 
 also maintaining the same performance is a significant barrier to the comm
 unity. Here, we discuss an approach for automating application and perform
 ance portability that is based on infrastructure as code provisioning of c
 loud clusters coupled with “compute environment as code” for configuring p
 ackage managers and containerized applications to fully utilize cloud reso
 urces. We benchmark two common MPI applications across different cloud ser
 vice providers. For each case, we present how some lessons learned and our
  best practices at several different levels within this HPC stack, includi
 ng the configuration of high performance networking and storage options, h
 as enabled both application and performance portability.\n\nTag: Cloud Com
 puting, Portability\n\nRegistration Category: Tech Program Reg Pass, Exhib
 its Reg Pass\n\nSession Chair: Junuh Kiang (Indiana University)\n\n
END:VEVENT
END:VCALENDAR