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:20240626T180034Z
LOCATION:Level 2 Lobby
DTSTART;TZID=America/Los_Angeles:20240626T180000
DTEND;TZID=America/Los_Angeles:20240626T190000
UID:dac_DAC 2024_sess237_RESEARCH1421@linklings.com
SUMMARY:A Fast IR-drop Modeling for In-RRAM Computing Considering Data All
 ocation
DESCRIPTION:Work-in-Progress Poster\n\nShih-Han Chang, Tung Lin, and Chien
 -Nan Liu (National Yang Ming Chiao Tung University)\n\nIn recent years, In
 -RRAM Computing (IRC) is a promising technique for deep neural network (DN
 N) applications. Combined with proper pruning technique, the cost and ener
 gy of DNN computation can be further reduced. However, IRC often suffers f
 rom various non-ideal effects in RRAM arrays such as the sneak path and IR
 -drop, which greatly affect the computation accuracy. Therefore, accurate 
 error injection is required in the verification at early design stage. Con
 ventional random disturbance and equation-based approaches did not conside
 r the data allocation issue, which may incur larger errors for the sparse 
 matrix generated by data pruning techniques. In this paper, a fast and acc
 urate IR-drop model is proposed to reflect the data-dependent effects, whi
 ch is able to offer accurate error injection in the DNN training phase wit
 h sparse matrix. As shown in the experimental results, the proposed model 
 shows a good match to the HSPICE results even if the data allocation becom
 es non-uniform. With the proposed simple model, the accuracy degradation o
 f real NN applications can be well observed, even for large RRAM arrays.\n
 \nTopic: AI, Autonomous Systems, Cloud, Design, EDA, Embedded Systems, IP,
  Security
END:VEVENT
END:VCALENDAR