BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Europe/Rome
BEGIN:DAYLIGHT
DTSTART:20260329T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20261025T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260424T195817Z
UID:26DE812F-7B05-4D2A-AC47-500CB393D81B
DTSTART;TZID=Europe/Rome:20260421T143000
DTEND;TZID=Europe/Rome:20260421T153000
DESCRIPTION:Distributed AI workloads are unique in the way they make use of
  the network. As a consequence\, traditional networking solutions are not 
 ideal for interconnecting the massive number of processors deployed in clu
 sters for AI training and inference.\n\nGiven that the network plays a key
  role in Distributed AI's performance and power consumption\, academia and
  industry have devoted an enormous amount effort to develop specialized so
 lutions. After analyzing the unique networking requirements of distributed
  AI workloads\, this talk describes the network architectures commonly use
 d to achieve the extreme scale of AI clusters and the standard protocol st
 acks developed for this specific context\, covering both scale scale out a
 nd scale out domains.\n\nCo-sponsored by: Internet Society Italy Chapter\;
  University of Pisa\n\nSpeaker(s): Mario Baldi\n\nRoom: Aula Pacinotti\, B
 ldg: Scuola di Ingegneria\, Largo L. Lazzarino 1\, Pisa\, Toscana\, Italy\
 , 56122
LOCATION:Room: Aula Pacinotti\, Bldg: Scuola di Ingegneria\, Largo L. Lazza
 rino 1\, Pisa\, Toscana\, Italy\, 56122
ORGANIZER:daniele.tarchi@unifi.it
SEQUENCE:22
SUMMARY:Networking for Distributed Artificial Intelligence
URL;VALUE=URI:https://events.vtools.ieee.org/m/553661
X-ALT-DESC:Description: <br /><p data-path-to-node="6">Distributed AI workl
 oads are unique in the way they make use of the network. As a consequence\
 , traditional networking solutions are not ideal for interconnecting the m
 assive number of processors deployed in clusters for AI training and infer
 ence.</p>\n<p data-path-to-node="7">Given that the network plays a key rol
 e in Distributed AI's performance and power consumption\, academia and ind
 ustry have devoted an enormous amount effort to develop specialized soluti
 ons. After analyzing the unique networking requirements of distributed AI 
 workloads\, this talk describes the network architectures commonly used to
  achieve the extreme scale of AI clusters and the standard protocol stacks
  developed for this specific context\, covering both scale scale out and s
 cale out domains.</p>
END:VEVENT
END:VCALENDAR