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:20260522T150118Z
LOCATION:D175
DTSTART;TZID=America/Chicago:20181112T163000
DTEND;TZID=America/Chicago:20181112T170000
UID:submissions.supercomputing.org_SC18_sess176_ws_llvmf110@linklings.com
SUMMARY:AIWC: OpenCL-Based Architecture Independent Workload Characterizat
 ion
DESCRIPTION:Beau Johnston and Josh Milthorpe (Australian National Universi
 ty)\n\nMeasuring performance-critical characteristics of application workl
 oads is important both for developers, who must understand and optimize th
 e performance of codes, as well as designers and integrators of HPC system
 s, who must ensure that compute architectures are suitable for the intende
 d workloads. However, if these workload characteristics are tied to archit
 ectural features that are specific to a particular system, they may not ge
 neralize well to alternative or future systems. An architecture-independen
 t method ensures an accurate characterization of inherent program behaviou
 r, without bias due to architecture-dependent features that vary widely be
 tween different types of accelerators. This work presents the first archit
 ecture-independent workload characterization framework for heterogeneous c
 ompute platforms, proposing a set of metrics determining the suitability a
 nd performance of an application on any parallel HPC architecture. The too
 l, AIWC, is a plugin for the open-source Oclgrind simulator. It supports p
 arallel workloads and is capable of characterizing OpenCL codes currently 
 in use in the supercomputing setting. AIWC simulates an OpenCL device by d
 irectly interpreting LLVM instructions, and the resulting metrics may be u
 sed for performance prediction and developer feedback to guide device-spec
 ific optimizations. An evaluation of the metrics collected over a subset o
 f the Extended OpenDwarfs Benchmark Suite is also presented.\n\nTag: Progr
 am Transformation, Programming Systems\n\nRegistration Category: Workshop 
 Reg Pass\n\n
END:VEVENT
END:VCALENDAR