BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20191206T142453Z
LOCATION:Substanz 1\, 2
DTSTART;TZID=Europe/Stockholm:20180626T083000
DTEND;TZID=Europe/Stockholm:20180626T100000
UID:isc_hpc_ISC High Performance 2018_sess113_post111@linklings.com
SUMMARY:(RP13) Cross-architectural Modelling of Power Consumption Using Ne
 ural Networks
DESCRIPTION:Research Poster\n\n(RP13) Cross-architectural Modelling of Pow
 er Consumption Using Neural Networks\n\nPuzovic, Lee, Elisseev\n\nOn the p
 ath to Exascale, goal of High Performance Computing (HPC) to achieve maxim
 um performance, becomes a goal of achieving maximum performance under stri
 ct power constraint. Novel approaches to hardware and software co-design o
 f modern HPC systems have to be developed to address such challenges. In t
 his paper we present Neural Networks-based model for estimating power cons
 umption on HPC systems from metrics obtained using hardware performance co
 unters across three different architectures, Intel 64, IBM Power and Caviu
 m ThunderX ARMv8. We demonstrate that our approach can provide 2 to 3 time
 s better accuracy predictions for power consumption compared to the curren
 t state of the art. Results of our work can be used as a starting point fo
 r developing unified, cross-architectural models for predicting power cons
 umption.\n\nPasses: Conference Pass, AI/Machine Learning/Deep Learning, Co
 mputer Architecture, Energy consumption, Performance Analysis and Optimiza
 tion, Performance Tools\n\nTag: Conference Pass, AI/Machine Learning/Deep 
 Learning, Computer Architecture, Energy consumption, Performance Analysis 
 and Optimization, Performance Tools
URL:https://2018.isc-program.com/presentation/?id=post111&sess=sess113
END:VEVENT
END:VCALENDAR