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:20260730T152640Z
LOCATION:Exhibit Hall
DTSTART;TZID=America/Los_Angeles:20260728T174100
DTEND;TZID=America/Los_Angeles:20260728T174100
UID:dac_DAC 2026_sess306_LBR146@linklings.com
SUMMARY:Polaris: Enabling PIM-Optimized Low-Energy SpGEMM using a Remote-A
 ccess-Aware Mapping Solution
DESCRIPTION:Helya Hosseini (University of Maryland, College Park); Christi
 na Giannoula (Max Planck Institute for Software Systems); and Bahar Asgari
  (University of Maryland, College Park)\n\nSparse kernels are widely used 
 in applications ranging from scientific computing to machine learning. Amo
 ng them, SpGEMM is particularly challenging due to irregular memory access
 es, making it highly memory-bound and dominated by data movement. Processi
 ng-in-Memory architectures can mitigate this cost by performing computatio
 n within memory banks, but irregular sparsity often causes load imbalance 
 and costly cross-bank accesses. To address these challenges, we propose Po
 laris, a PIM-aware map-\nping solution for efficient SpGEMM on HBM-based s
 ystems. Polaris leverages a row-wise product formulation and a sparsity-aw
 are mapping strategy to improve locality and balance workloads across bank
 s, reducing data movement and achieving on average, 2× lower energy consum
 ption and up to 8× speedup over prior designs.\n\nTrack: Student\n\n
END:VEVENT
END:VCALENDAR