BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000711Z
LOCATION:401-402
DTSTART;TZID=America/Denver:20231116T113000
DTEND;TZID=America/Denver:20231116T120000
UID:submissions.supercomputing.org_SC23_sess158_pap376@linklings.com
SUMMARY:ReFloat: Low-Cost Floating-Point Processing in ReRAM for Accelerat
 ing Iterative Linear Solvers
DESCRIPTION:Linghao Song (University of California Los Angeles (UCLA)), Fa
 n Chen (Indiana University), and Hai Li and Yiran Chen (Duke University)\n
 \nResistive random access memory (ReRAM) is a promising technology that ca
 n perform low-cost and in-situ matrix-vector multiplication (MVM) in analo
 g domain. Scientific computing requires high-precision floating-point (FP)
  processing. However, performing floating-point computation in ReRAM is ch
 allenging because of high hardware cost and execution time due to the larg
 e FP value range. In this work we present ReFloat, a data format and an ac
 celerator architecture, for low-cost and high-performance floating-point p
 rocessing in ReRAM for iterative linear solvers. ReFloat matches the ReRAM
  crossbar hardware and represents a block of FP values with reduced bits a
 nd an optimized exponent base for a high range of dynamic representation. 
 Thus, ReFloat achieves less ReRAM crossbar consumption and fewer processin
 g cycles and overcomes the noncovergence issue in a prior work. The evalua
 tion on the SuiteSparse matrices shows that ReFloat achieves 5.02x to 84.2
 8x improvement in terms of solver time compared to a state-of-the-art ReRA
 M based accelerator.\n\nTag: Algorithms, Linear Algebra, Post-Moore Comput
 ing\n\nRegistration Category: Tech Program Reg Pass\n\nReproducibility Bad
 ges: Artifact Available, Artifact Functional, Results Reproduced\n\nSessio
 n Chair: Julien Langou (University of Colorado Denver; University of Color
 ado, Department of Mathematical and Statistical Sciences)\n\n
END:VEVENT
END:VCALENDAR