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:20260402T024534Z
LOCATION:Level 3 Lobby
DTSTART;TZID=America/Los_Angeles:20250622T180000
DTEND;TZID=America/Los_Angeles:20250622T190000
UID:dac_DAC 2025_sess261_RESEARCH897@linklings.com
SUMMARY:GRL: Redesign Distributed Reinforcement Learning Training on One G
 PU
DESCRIPTION:Zhikuang Xin, Zhenghong Wu, Rongqiang Cao, haoyu Wang, Haisha 
 Zhao, Jue Wang, and Yangang Wang (University of Chinese Academy of Science
 s)\n\nReinforcement learning is computationally intensive due to frequent 
 data exchanges between learners and actors., making it hard to fully utili
 ze the GPU. To address this, we propose a RL framework GRL, marking the fi
 rst time the complete RL process is deployed on one GPU. Based on the feat
 ures of GPU, we design the lock-free model queue and the fused actors to e
 nhance the experience throughput of framework. We propose an auto-configur
 ator to adjust the runtime configuration to speed up the whole framework. 
 We test GRL in various RL environments. GRL achieves an improvement on thr
 oughput from 4 to 200 times.\n\nTracks: DES5: Emerging Device and Intercon
 nect Technologies\n\n
END:VEVENT
END:VCALENDAR