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:20260522T150110Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181114T083000
DTEND;TZID=America/Chicago:20181114T170000
UID:submissions.supercomputing.org_SC18_sess323_post256@linklings.com
SUMMARY:GPGPU Performance Estimation with Core and Memory Frequency Scalin
 g
DESCRIPTION:Qiang Wang and Xiaowen Chu (Hong Kong Baptist University)\n\nG
 raphics processing units (GPUs) support dynamic voltage and frequency scal
 ing to balance computational performance and energy consumption. However, 
 simple and accurate performance estimation for a given GPU kernel under di
 fferent frequency settings is still lacking for real hardware, which is im
 portant to decide the best frequency configuration for energy saving. We r
 eveal a fine-grained analytical model to estimate the execution time of GP
 U kernels with both core and memory frequency scaling. Over a wide scaling
  range of both core and memory frequencies among 20 GPU kernels, our model
  achieves accurate results (4.83% error on average) with real hardware. Co
 mpared to the cycle-level simulators, our model only needs simple micro-be
 nchmarks to extract a set of hardware parameters and kernel performance co
 unters to produce such high accuracy.\n\nRegistration Category: Tech Progr
 am Reg Pass, Exhibits Reg Pass\n\n
END:VEVENT
END:VCALENDAR