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:20240626T180033Z
LOCATION:3012\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240625T163000
DTEND;TZID=America/Los_Angeles:20240625T164500
UID:dac_DAC 2024_sess112_RESEARCH2055@linklings.com
SUMMARY:FHE-CGRA: Enable Efficient Acceleration of Fully Homomorphic Encry
 ption on CGRAs
DESCRIPTION:Research Manuscript\n\nMiaomiao Jiang, Yilan Zhu, and Honghui 
 You (Shandong University); Cheng Tan (Google); Zhaoying Li (National Unive
 rsity of Singapore); Jiming Xu (Ant Group); and Lei Ju (Shandong Universit
 y)\n\nFully Homomorphic Encryption (FHE) is a privacy-preserving technique
  that allows computation directly on encrypted data. In this work, we inve
 stigate execution the of FHE machine learning (ML) applications. We show t
 hat the runtime hardware reconfigurability of the underlying execution uni
 ts of homomorphic operations is highly desirable for efficient hardware re
 source utilization. Based on the observation, we propose FHE-CGRA, a coars
 e-grained reconfigurable architecture (CGRA) acceleration framework for en
 d-to-end homomorphic applications. The experiment shows that FHE-CGRA achi
 eves up to 8.15x speedup against a conventional CGRA for accelerating FHE-
 encrypted convolution neural network (FHE-CNN) models, and 16.48x power ef
 ficiency w.r.t. the state-of-the-art FPGA.\n\nTopic: Design\n\nKeyword: So
 C, Heterogeneous, and Reconfigurable Architectures\n\nSession Chairs: Dimi
 trios Soudris (National Technical University of Athens) and George Tzimpra
 gos (University of Michigan)
END:VEVENT
END:VCALENDAR