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:20181221T160905Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181114T083000
DTEND;TZID=America/Chicago:20181114T170000
UID:submissions.supercomputing.org_SC18_sess323_post118@linklings.com
SUMMARY:Portable Parallel Performance via Multi-Dimensional Homomorphisms
DESCRIPTION:Poster\nTech Program Reg Pass, Exhibits Reg Pass\n\nPortable P
 arallel Performance via Multi-Dimensional Homomorphisms\n\nRasch, Schulze,
  Gorlatch\n\nAchieving portable performance over different parallel archit
 ectures and varying problem sizes is hard: e.g., a program optimized for m
 ulti-core CPUs on large input sizes can significantly differ from the same
  program optimized for Graphics Processing Units (GPUs) on small sizes.\n\
 nWe propose an approach to ensuring portability of performance by relying 
 on multi-dimensional homomorphisms (MDHs) -- a class of parallelizable fun
 ctions that cover important application areas including linear algebra rou
 tines (BLAS) and stencil computations. We develop an extended OpenCL imple
 mentation schema for MDHs that is generic in the performance-critical para
 meters of the OpenCL model, and we enable portability of performance by be
 ing automatically optimized for different target architectures and input s
 izes using the auto-tuning approach. \n\nOur results demonstrate competiti
 ve and often even significantly better performance than state-of-the-art a
 pproaches for BLAS and Stencil as used in the important application area o
 f deep learning.
URL:https://sc18.supercomputing.org/presentation/?id=post118&sess=sess323
END:VEVENT
END:VCALENDAR