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:20211115T091000
DTEND;TZID=America/Chicago:20211115T093000
UID:submissions.supercomputing.org_SC21_sess422_ws_pmbss106@linklings.com
SUMMARY:Memory Demands in Disaggregated HPC: How Accurate Do We Need to Be
 ?
DESCRIPTION:Workshop\n\nMemory Demands in Disaggregated HPC: How Accurate 
 Do We Need to Be?\n\nVieira Zacarias, Petrucci, Carpenter\n\nJobs running 
 on HPC systems can vary dramatically due to the intrinsic differences in a
 pplication resource requirements (e.g. memory or cores). Since HPC applica
 tions run on a number of self-contained servers whose capacities are fixed
  at design time, there is often a mismatch between the resource provisioni
 ng and the needs of the submitted jobs, leading to stranded and underutili
 zed resources. This is because HPC systems assume the prevalent server-bas
 ed architecture, which couples together memory and processing resources wi
 thin a server. To cope efficiently with the demands, disaggregated memory 
 has been proposed to allow flexible and fine-grained allocation of memory 
 capacity to compute jobs.\n	\nThis paper makes an important step towards u
 nderstanding the facets of resource allocation and job requirements on dis
 aggregated memory systems. We analyze the implications on HPC system opera
 tion, user experience and system performance when resources can be overest
 imated by users. To conduct our studies, we leverage a disaggregated simul
 ation infrastructure implemented on a popular HPC resource manager. Our re
 sults show that the effects of doubling the memory demand in response time
  can be less than 8%.\n\nTag: Online Only, Accelerator-based Architectures
 , Applications, Computational Science, Emerging Technologies, Extreme Scal
 e Computing, File Systems and I/O, Heterogeneous Systems, Parallel Program
 ming Languages and Models, Performance, Scientific Computing, Software Eng
 ineering\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR