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:20260522T150110Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181114T083000
DTEND;TZID=America/Chicago:20181114T170000
UID:submissions.supercomputing.org_SC18_sess323_post219@linklings.com
SUMMARY:Automatic Generation of Mixed-Precision Programs
DESCRIPTION:Logan Moody (Lawrence Livermore National Laboratory, James Mad
 ison University); Nathan Pinnow (Lawrence Livermore National Laboratory, W
 estern Washington University); Michael O. Lam (James Madison University, L
 awrence Livermore National Laboratory); Harshitha Menon, Markus Schordan, 
 and G. Scott Lloyd (Lawrence Livermore National Laboratory); and Tanzima I
 slam (Western Washington University)\n\nFloating-point arithmetic is found
 ational to scientific computing in HPC, and choices about floating-point p
 recision can have a significant effect on the accuracy and speed of HPC co
 des. Unfortunately, current precision optimization tools require significa
 nt user interaction, and few work on the scale of HPC codes due to signifi
 cant analysis overhead. We propose an automatic search and replacement sys
 tem that finds the maximum speedup using mixed precision given a required 
 level of accuracy. To achieve this, we integrated three existing analysis 
 tools into a system that requires minimal input from the user. If a speedu
 p is found, our system can provide a ready-to-compile mixed-precision vers
 ion of the original program.\n\nRegistration Category: Tech Program Reg Pa
 ss, Exhibits Reg Pass\n\n
END:VEVENT
END:VCALENDAR