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:20220715T000417Z
LOCATION:2012\, Level 2
DTSTART;TZID=America/Los_Angeles:20220712T141500
DTEND;TZID=America/Los_Angeles:20220712T143000
UID:dac_DAC 2022_sess172_IP112@linklings.com
SUMMARY:Shared Virtual Output Queuing: Maximizing NoC Switch Utilization
DESCRIPTION:Engineering Tracks, IP\n\nShared Virtual Output Queuing: Maxim
 izing NoC Switch Utilization\n\nJo, Kim\n\nModern mobile SoC integrates th
 roughput-oriented processors such as GPU/NPU/ISP for interactive gaming an
 d AI camera experiences, which requires increasingly higher memory and NoC
  bandwidth for high-throughput computation. The switch contention by Head-
 of-Line (HoL) blocking is the major bottleneck limiting NoC bandwidth in m
 emory channel-interleaved systems. In order to resolve the HoL blocking an
 d maximize switch utilization, we evaluate the previous proposals and thei
 r limitations, and propose Shared VOQ enabling to 1) share a single buffer
  with packets heading to different destinations and 2) consider all matchi
 ng availability between sources and destinations to maximize switch utiliz
 ation. Our evaluation shows that Shared VOQ achieves up to 10.3% throughpu
 t improvement comparing to the conventional queue with up to 14.6% area ov
 erhead and negligible timing impact. It is also showed that achievable thr
 oughput by shared VOQ can be nearly reached to theoretical maximum.\n\nTop
 ic: IP
END:VEVENT
END:VCALENDAR