BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Brisbane
X-LIC-LOCATION:Australia/Brisbane
BEGIN:STANDARD
DTSTART:19920301T030000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20191203T080637Z
LOCATION:Great Hall 1&2
DTSTART;TZID=Australia/Brisbane:20191117T190800
DTEND;TZID=Australia/Brisbane:20191117T190900
UID:siggraphasia_SIGGRAPH Asia 2019_sess222_papers_318@linklings.com
SUMMARY:Automatically Translating Image Processing Libraries to Halide
DESCRIPTION:Technical Papers Fast-Forward \n\nAutomatically Translating Im
 age Processing Libraries to Halide\n\nAhmad, Ragan-Kelley, Cheung, Kamil\n
 \nThis paper presents Dexter, a new tool that automatically translates ima
 ge processing functions from a low-level general-purpose language to a hig
 h-level domain-specific language (DSL), allowing them to leverage cross-pl
 atform optimizations enabled by DSLs. Rather than building a classical syn
 tax-driven compiler to do this translation, Dexter leverages recent advanc
 es in program synthesis and program verification, along with a new domain-
 specific synthesis algorithm, to translate C++ image processing code to th
 e Halide DSL, while guaranteeing semantic equivalence. This new synthesis 
 algorithm scales and generalizes to much larger and more complex functions
  than prior work, including the ability to handle tiling, conditionals, an
 d multi-stage pipelines in the original low-level code. To demonstrate the
  effectiveness of our approach, we evaluate Dexter using real-world image 
 processing functions from Adobe Photoshop, a widely-used multi-platform im
 age processing program. Our results show that Dexter can translate 264 out
  of 353 functions in our test set, with the original implementations rangi
 ng from 20 to 150 lines of code. By leveraging Halide’s advanced auto-sche
 duling capabilities, we demonstrate median speedups of 7.03x and 4.52x for
  Dexter-translated functions as compared to the original implementations o
 n Intel and ARM architectures, respectively.\n\nRegistration Category: Pla
 tinum Pass, Full Conference Pass, Full Conference One-Day Pass, Basic Conf
 erence Pass, Student One-Day Pass, Exhibitor Pass
URL:https://sa2019.conference-program.com/presentation?id=papers_318&sess=
 sess222
END:VEVENT
END:VCALENDAR