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:20211207T054928Z
LOCATION:222
DTSTART;TZID=America/Chicago:20211118T110000
DTEND;TZID=America/Chicago:20211118T113000
UID:submissions.supercomputing.org_SC21_sess259_drs105@linklings.com
SUMMARY:Program Transformation for Automatic GPU-Offloading with OpenMP
DESCRIPTION:Doctoral Showcase, Posters\n\nProgram Transformation for Autom
 atic GPU-Offloading with OpenMP\n\nMishra, Chapman, Malik\n\nThanks to its
  ability to manage large data parallelism with low power consumption, hete
 rogeneous GPU computing has risen in the past decade. However, writing an 
 application for GPUs is an intensive manual effort that may include re-eng
 ineering data structures, as well as modifying large regions of code, to m
 ake effective use of the GPU’s computational power while keeping ove
 rheads moderate. Directive based programming models, such as OpenMP, are a
 n attractive approach due to their productivity benefits. Yet, programmers
  using OpenMP confront various hurdles, including programmability, data ha
 ndling, and parallelism. In this research, we design and build a compiler 
 framework that can automatically discover OpenMP kernels, recommend severa
 l potential OpenMP variants for offloading that kernel to the GPU, and usi
 ng a novel static neural network-based compile time cost model, predict an
 d return the most optimal of those variants. We divide our framework into 
 3 modules, each of which functions independently. Module 1 detects and ana
 lyzes an OpenMP kernel and suggests several variants, by applying various 
 potential code level transformations, for offloading that kernel to a GPU.
  In module 2, we define COMPOFF, which employs ML techniques (for the firs
 t time in OpenMP) to predict the Cost of OpenMP OFFloading statically. In 
 module 3, we modify the original source code using the analysis and predic
 tion from the other modules to modify the source code and returns newly ge
 nerated code that supports GPU offloading. Our preliminary findings indica
 te that our framework will aid scientists transfer their programs to the n
 ew heterogeneous computing environment.\n\nRegistration Category: Tech Pro
 gram Reg Pass
END:VEVENT
END:VCALENDAR