BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000712Z
LOCATION:405-406-407
DTSTART;TZID=America/Denver:20231115T163000
DTEND;TZID=America/Denver:20231115T170000
UID:submissions.supercomputing.org_SC23_sess183_pap263@linklings.com
SUMMARY:SYnergy: Fine-Grained Energy-Efficient Heterogeneous Computing for
  Scalable Energy Saving
DESCRIPTION:Kaijie Fan (TU Berlin, University of Salerno); Marco D'Antonio
 , Lorenzo Carpentieri, and Biagio Cosenza (University of Salerno); and Fed
 erico Ficarelli and Daniele Cesarini (CINECA)\n\nEnergy-efficient computin
 g uses power management techniques such as frequency scaling to save energ
 y. Implementing energy-efficient techniques on large-scale computing syste
 ms is challenging. While most modern architectures, including GPUs, are ca
 pable of frequency scaling, these features are often not available on larg
 e systems.  \n\nWe propose SYnergy, a novel energy-efficient approach that
  spans languages, compilers, runtimes, and job schedulers to achieve unpre
 cedented fine-grained energy savings on large-scale heterogeneous clusters
 . SYnergy defines an extension to the SYCL programming model that allows p
 rogrammers to define a specific energy goal for each kernel.  Through comp
 iler integration and a machine learning model, each kernel is statically o
 ptimized for the specific target. The methodology is inherently portable a
 nd has been evaluated on both NVIDIA and AMD GPUs. Experimental results sh
 ow unprecedented improvements in energy and energy-related metrics on real
 -world applications, as well as scalable energy savings on a 64-GPU cluste
 r.\n\nTag: Cloud Computing, Distributed Computing, Energy Efficiency, Perf
 ormance Measurement, Modeling, and Tools\n\nRegistration Category: Tech Pr
 ogram Reg Pass\n\nReproducibility Badges: Artifact Available, Artifact Fun
 ctional, Results Reproduced\n\nSession Chair: Radu Prodan (University of I
 nnsbruck, Austria)\n\n
END:VEVENT
END:VCALENDAR