BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230124T171519Z
LOCATION:C146
DTSTART;TZID=America/Chicago:20221113T141500
DTEND;TZID=America/Chicago:20221113T144000
UID:submissions.supercomputing.org_SC22_sess424_ws_scalah101@linklings.com
SUMMARY:Threshold Pivoting for Dense LU Factorization
DESCRIPTION:Workshop\n\nThreshold Pivoting for Dense LU Factorization\n\nL
 indquist, Gates, Luszczek, Dongarra\n\nLU factorization is a key approach 
 for solving large, dense systems of linear equations. Partial row pivoting
  is commonly used to ensure numerical stability; however, the data movemen
 t needed for the row interchanges can reduce performance. To improve this,
  we propose using threshold pivoting to find pivots almost as good as thos
 e selected by partial pivoting but that result in less data movement. Our 
 theoretical analysis bounds the element growth similarly to partial pivoti
 ng; however, it also shows that the growth of threshold pivoting for a giv
 en matrix cannot be bounded by that of partial pivoting and vice versa. Ad
 ditionally, we experimentally tested the approach on the Summit supercompu
 ter. Threshold pivoting improved performance by up to 32% without a signif
 icant effect on accuracy. For a more aggressive configuration with up to o
 ne digit of accuracy lost, the improvement was as high as 44%.\n\nSession 
 Format: Recorded\n\nTag: Algorithms, Exascale Computing, Extreme Scale Com
 puting, Heterogeneous Systems, Post-Moore Computing, Quantum Computing\n\n
 Registration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR