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:20181221T160731Z
LOCATION:D174
DTSTART;TZID=America/Chicago:20181116T090000
DTEND;TZID=America/Chicago:20181116T092000
UID:submissions.supercomputing.org_SC18_sess146_ws_ftxs103@linklings.com
SUMMARY:Fault Tolerant Cholesky Factorization on GPUs
DESCRIPTION:Workshop\nResiliency, Scientific Computing, Workshop Reg Pass\
 n\nFault Tolerant Cholesky Factorization on GPUs\n\nLoh, Saluja, Ramanatha
 n\n\nDirect Cholesky-based solvers are typically used to solve large linea
 r systems where the coefficient matrix is symmetric positive definite. The
 se solvers offer faster performance in solving such linear systems, compar
 ed to other more general solvers such as LU and QR solvers. In recent days
 , graphics processing units (GPUs) have become a popular platform for scie
 ntific computing applications, and are increasingly being used as major co
 mputational units in supercomputers. However, GPUs are susceptible to tran
 sient faults caused by events such as alpha particle strikes and power flu
 ctuations. As a result, the possibility of an error increases as more and 
 more GPU computing nodes are used. In this paper, we introduce two efficie
 nt fault tolerance schemes for the Cholesky factorization method, and stud
 y their performance using a direct Cholesky solver in the presence of faul
 ts. We utilize a transient fault injection mechanism for NVIDIA GPUs and c
 ompare our schemes with a traditional checksum fault tolerance technique, 
 and show that our proposed schemes have superior performance, good error c
 overage and low overhead.
URL:https://sc18.supercomputing.org/presentation/?id=ws_ftxs103&sess=sess1
 46
END:VEVENT
END:VCALENDAR

