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:3001\, Level 3
DTSTART;TZID=America/Los_Angeles:20250624T161500
DTEND;TZID=America/Los_Angeles:20250624T163000
UID:dac_DAC 2025_sess121_RESEARCH847@linklings.com
SUMMARY:3D-SubG: A 3D Stacked Hybrid Processing Near/In-Memory Accelerator
  for Subgraph GNNs
DESCRIPTION:Guoxiang Li, Runnan Xu, and Ruohang Xu (Peking University); Yi
 kan Qiu (Hangzhou Hikvision Digital Technology Co., Ltd.); and Renati Tuer
 hong, Muhan Zhang, Le Ye, and Yufei Ma (Peking University)\n\nSubgraph Gra
 ph Neural Networks (GNNs) are emerging as a promising approach to enhance 
 GNN expressiveness, but their more complex graph structures with numerous 
 independent and irregular subgraphs pose significant hardware deployment c
 hallenges. In this work, we propose 3D-SubG, a 3D stacked hybrid processin
 g-near/in-memory accelerator for subgraph GNNs. With hybrid bonding packag
 ing technology, a logic die is 3D stacked with a DRAM die for highly paral
 lel memory accesses. The logic die employs digital SRAM-based processing-i
 n-memory (PIM) macros to boost computation density and minimize data trans
 fer. We further propose a bit-level non-zero gathering method to exploit g
 raph sparsity for PIM, a workload-balanced mapping strategy for subgraph a
 llocation onto different logic-to-DRAM blocks, and a distributed global po
 oling approach to reduce inter-block data movements. Experimental results 
 show that 3D-SubG achieves average improvements of 146.11× in performance,
  934.18× in area efficiency, and 1171.80× in energy efficiency compared to
  RTX 3090Ti.\n\nTopics: Design\n\nTracks: DES2B: In-memory and Near-memory
  Computing Architectures, Applications and Systems\n\nSession Chairs: Arma
 n Roohi (University of Illinois, Chicago) and Abhronil Sengupta (Pennsylva
 nia State University)\n\n
END:VEVENT
END:VCALENDAR