BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000712Z
LOCATION:702
DTSTART;TZID=America/Denver:20231112T172000
DTEND;TZID=America/Denver:20231112T173000
UID:submissions.supercomputing.org_SC23_sess436_ws_ia104@linklings.com
SUMMARY:Accelerating Deep Neural Network Guided MCTS Using Adaptive Parall
 elism
DESCRIPTION:Yuan Meng, Qian Wang, Tianxin Zu, and Viktor Prasanna (Univers
 ity of Southern California (USC))\n\nDeep Neural Network guided Monte-Carl
 o Tree Search (DNN-MCTS) is a powerful class of AI algorithms. The DNN ope
 rations are highly parallelizable, but tree search operations are sequenti
 al and often become the system bottleneck.  Existing MCTS parallel schemes
  on CPU platforms either exploit data parallelism but sacrifice memory acc
 ess latency, or take advantage of local cache for low-latency accesses but
  constrain the search to a single thread.  This work analyzes the tradeoff
  of these parallel schemes, and proposes an adaptive parallel scheme that 
 optimally chooses the MCTS component's parallel scheme on the CPU. Additio
 nally, an efficient method for searching the optimal communication batch s
 ize when the CPU interfaces with DNN operations on an accelerator(GPU) is 
 proposed.  Using a DNN-MCTS algorithm on board game benchmarks, we show th
 at our work is able to adaptively generate the best-performing parallel im
 plementation, leading to a range of 1.5-3 times speedup compared with the 
 baseline methods.\n\nTag: Algorithms, Applications, Architecture and Netwo
 rks\n\nRegistration Category: Workshop Reg Pass\n\nSession Chairs: Sanjukt
 a Bhowmick (University of North Texas), Vito Giovanni Castellana (Pacific 
 Northwest National Laboratory (PNNL)), John Feo (Pacific Northwest Nationa
 l Laboratory (PNNL)), Marco Minutoli (Advanced Micro Devices), and Antonin
 o Tumeo (Pacific Northwest National Laboratory (PNNL))\n\n
END:VEVENT
END:VCALENDAR