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:20181221T160904Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181113T083000
DTEND;TZID=America/Chicago:20181113T170000
UID:submissions.supercomputing.org_SC18_sess322_post256@linklings.com
SUMMARY:GPGPU Performance Estimation with Core and Memory Frequency Scalin
 g
DESCRIPTION:Poster\nTech Program Reg Pass, Exhibits Reg Pass\n\nGPGPU Perf
 ormance Estimation with Core and Memory Frequency Scaling\n\nWang, Chu\n\n
 Graphics processing units (GPUs) support dynamic voltage and frequency sca
 ling to balance computational performance and energy consumption. However,
  simple and accurate performance estimation for a given GPU kernel under d
 ifferent frequency settings is still lacking for real hardware, which is i
 mportant to decide the best frequency configuration for energy saving. We 
 reveal a fine-grained analytical model to estimate the execution time of G
 PU kernels with both core and memory frequency scaling. Over a wide scalin
 g range of both core and memory frequencies among 20 GPU kernels, our mode
 l achieves accurate results (4.83% error on average) with real hardware. C
 ompared to the cycle-level simulators, our model only needs simple micro-b
 enchmarks to extract a set of hardware parameters and kernel performance c
 ounters to produce such high accuracy.
URL:https://sc18.supercomputing.org/presentation/?id=post256&sess=sess322
END:VEVENT
END:VCALENDAR