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:20260522T150120Z
LOCATION:D171/173
DTSTART;TZID=America/Chicago:20181116T085500
DTEND;TZID=America/Chicago:20181116T091000
UID:submissions.supercomputing.org_SC18_sess145_ws_p3hpc107@linklings.com
SUMMARY:An Empirical Roofline Methodology for Quantitatively Assessing Per
 formance Portability
DESCRIPTION:Charlene Yang, Rahulkumar Gayatri, Thorsten Kurth, Protonu Bas
 u, and Zahra Ronaghi (Lawrence Berkeley National Laboratory); Adedoyin Ade
 tokunbo (Los Alamos National Laboratory); and Brian Friesen, Brandon Cook,
  Douglas Doerfler, Leonid Oliker, Jack Deslippe, and Samuel Williams (Lawr
 ence Berkeley National Laboratory)\n\nSystem and node architectures contin
 ue to diversify to better balance on-node computation, memory capacity, me
 mory bandwidth, interconnect bandwidth, power, and cost for specific compu
 tational workloads. For many applications developers, however, achieving p
 erformance portability (effectively exploiting the capabilities of multipl
 e architectures) is a desired goal. Unfortunately, dramatically different 
 per-node performance coupled with differences in machine balance can lead 
 to developers being unable to determine whether they have attained perform
 ance portability or simply written portable code. The Roofline model provi
 des a means of quantitatively assessing how well a given application makes
  use of a target platform's computational capabilities. In this paper, we 
 extend the Roofline model so that it 1) empirically captures a more realis
 tic set of performance bounds for CPUs and GPUs,  2) factors in the true c
 ost of different floating-point instructions when counting FLOPs, 3) incor
 porates the effects of different memory access patterns, and 4) with appro
 priate pairing of code performance and Roofline ceiling, facilitates the p
 erformance portability analysis.\n\nTag: Heterogeneous Systems, Performanc
 e\n\nRegistration Category: Workshop Reg Pass\n\nSession Chair: Rob Neely 
 (Lawrence Livermore National Laboratory)\n\n
END:VEVENT
END:VCALENDAR