BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20211208T164407Z
LOCATION:3016
DTSTART;TZID=America/Los_Angeles:20211207T135200
DTEND;TZID=America/Los_Angeles:20211207T141500
UID:dac_DAC 2021_sess126_RESEARCH1267@linklings.com
SUMMARY:ST2GPU: An Energy-Efficient GPU Design with Spatio-Temporal Shared
 -Thread Speculative Adders
DESCRIPTION:Research Manuscript\n\nST2GPU: An Energy-Efficient GPU Design 
 with Spatio-Temporal Shared-Thread Speculative Adders\n\nKandiah\n\nModern
  GPUs employ thousands of cores, yielding higher performance but also high
 er power consumption. To meet performance targets while staying within a r
 easonable power budget, designers have to make these execution cores incre
 asingly more power efficient. One way to increase their power efficiency i
 s to employ power-efficient adders. We observe that consecutive arithmetic
  computations from the same code location are highly correlated and propos
 e ST2GPU, a GPU architecture that uses history-based speculative adders th
 at produce guaranteed correct results while saving 70% of the nominal adde
 r power. We estimate that ST2GPU saves 21% chip energy with negligible per
 formance and area overheads.\n\nTopic: EDA\n\nKeyword: Cross-Layer Power A
 nalysis and Low-Power Design\n\nVirtual Program: Presented In-Person
END:VEVENT
END:VCALENDAR
