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:20211208T164410Z
LOCATION:3016
DTSTART;TZID=America/Los_Angeles:20211208T142200
DTEND;TZID=America/Los_Angeles:20211208T145200
UID:dac_DAC 2021_sess113_RESEARCH439@linklings.com
SUMMARY:F-CAD: A Framework to Explore Hardware Accelerators for Codec Avat
 ar Decoding
DESCRIPTION:Research Manuscript\n\nF-CAD: A Framework to Explore Hardware 
 Accelerators for Codec Avatar Decoding\n\nZhang\n\nExisting hardware accel
 erators fail to deliver sufficient performance and efficiency for VR codec
  avatar decoding with multi-branch DNN inference workloads. To address thi
 s problem, we propose F-CAD, an automation framework to explore and delive
 r optimized architectures for codec avatar decoding given application-spec
 ific and hardware resource constraints. Novel technologies include 1) a ne
 w accelerator architecture to efficiently handle multi-branch DNNs; 2) a h
 igh-dimensional multi-branch design space to enable fine-grained architect
 ure configurations; 3) an effective architecture exploration for picking t
 he optimized hardware design. F-CAD is the first automation tool that supp
 orts accelerator design, exploration, and evaluation for codec avatar deco
 ding.\n\nTopic: Design\n\nKeyword: AI/ML System Design\n\nVirtual Program:
  Presented In-Person
END:VEVENT
END:VCALENDAR
