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:20181221T160731Z
LOCATION:C141/143/149
DTSTART;TZID=America/Chicago:20181114T163000
DTEND;TZID=America/Chicago:20181114T170000
UID:submissions.supercomputing.org_SC18_sess215_pap503@linklings.com
SUMMARY:ADAPT: Algorithmic Differentiation Applied to Floating-Point Preci
 sion Tuning
DESCRIPTION:Paper\nAlgorithms, Applications, Architectures, Compiler Analy
 sis and Optimization, Floating Point, Performance, Precision, Programming 
 Systems, Tools, Tech Program Reg Pass\n\nADAPT: Algorithmic Differentiatio
 n Applied to Floating-Point Precision Tuning\n\nMenon, Lam, Osei-Kuffuor, 
 Schordan, Lloyd...\n\nHPC applications extensively use floating point arit
 hmetic operations to solve computational problems in various domains. Mixe
 d precision computing, use of lowest precision data type sufficient to ach
 ieve a desired accuracy, have been explored to improve performance, reduce
  power consumption and data movement. Manually optimizing the program to u
 se mixed precision is challenging. In this work, we present ADAPT, an appr
 oach for mixed precision analysis on HPC workloads while providing guarant
 ees about the final output error. Our approach uses algorithmic differenti
 ation to accurately estimate the output error for mixed precision configur
 ation. ADAPT provides floating-point precision sensitivity of programs, wh
 ich highlights regions of the code that that can potentially be converted 
 to lower precision, is used to make algorithmic choices and develop mixed 
 precision configurations. We evaluate ADAPT on six benchmarks and a proxy 
 application and show that we are able to achieve a speedup of 1.2x on the 
 proxy application, LULESH.
URL:https://sc18.supercomputing.org/presentation/?id=pap503&sess=sess215
END:VEVENT
END:VCALENDAR