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:20211207T055347Z
LOCATION:Online
DTSTART;TZID=America/Chicago:20211115T110000
DTEND;TZID=America/Chicago:20211115T113000
UID:submissions.supercomputing.org_SC21_sess422_ws_pmbsf122@linklings.com
SUMMARY:An Extended Roofline Performance Model with PCI-E and Network Ceil
 ings
DESCRIPTION:Workshop\n\nAn Extended Roofline Performance Model with PCI-E 
 and Network Ceilings\n\nDufek, Deslippe, Lin, Yang, Cook...\n\nIn this wor
 k, we evaluate the utility of adding two new diagonal ceilings to the roof
 line model related to PCI-E and effective network bandwidths to provide in
 sights into how communication impacts the performance of large-scale paral
 lel applications. The roofline performance analysis is based on two benchm
 ark problems: scalar dense matrix addition and dense symmetric eigen-probl
 em with complex matrix. The experiments were conducted on the NERSC Cori s
 upercomputer at Lawrence Berkeley National Laboratory, on both the CPU-onl
 y and CPU+GPU compute nodes. The study reveals the value of incorporating 
 these two new ceilings into the roofline model, in addition to the existin
 g memory bandwidth and compute ceilings, in order to ease the identificati
 on of performance bottlenecks to better guide the performance optimization
  process, particularly in the limit of diminishing strong and weak scaling
 . We highlight the importance of comparing obtained application roofline p
 oints to customized ceilings for the communication and data-access pattern
 s present. In this way, the effects of both throughput and latency can be 
 captured in the model.\n\nTag: Online Only, Accelerator-based Architecture
 s, Applications, Computational Science, Emerging Technologies, Extreme Sca
 le Computing, File Systems and I/O, Heterogeneous Systems, Parallel Progra
 mming Languages and Models, Performance, Scientific Computing, Software En
 gineering\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR
