BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20241120T082409Z
LOCATION:HG F 3
DTSTART;TZID=Europe/Stockholm:20240604T160000
DTEND;TZID=Europe/Stockholm:20240604T163000
UID:submissions.pasc-conference.org_PASC24_sess112_msa260@linklings.com
SUMMARY:Machine Learning for Performance Engineering Across Applications
DESCRIPTION:Minisymposium\n\nTal Ben-Nun (Lawrence Livermore National Labo
 ratory)\n\nDeveloping fast and portable HPC codes is an evolving process t
 hat spans the entire lifetime of an application, adapting to changes in ta
 rget hardware and new software optimization techniques over time. Unfortun
 ately, because of subtle differences in codes, the lessons learned from op
 timizing one application are often re-learned in others, despite sharing s
 imilar optimization spaces. This talk will shine a spotlight on ML-based t
 echniques to identify optimization opportunities across applications. Usin
 g fuzzy matching, performance embedding spaces, and transfer tuning, perfo
 rmance engineers are now able to cluster subprograms by static and dynamic
  characteristics.\nWe will discuss the productivity benefits of state-of-t
 he-art methods and the potential computational reduction (within and acros
 s applications) - turning hundreds of hours of auto-tuning to a "performan
 ce database" query and local search - using a real-world case study in an 
 atmospheric model. Lastly, we will discuss future directions in the field 
 and how we can potentially leverage machine learning to aid performance en
 gineering more generally.\n\nDomain: Computational Methods and Applied Mat
 hematics\n\nSession Chairs: Florina Ciorba (University of Basel) and Harsh
 itha Menon (Lawrence Livermore National Laboratory)
END:VEVENT
END:VCALENDAR
