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:20200129T163556Z
LOCATION:503-504
DTSTART;TZID=America/Denver:20191122T090700
DTEND;TZID=America/Denver:20191122T092500
UID:submissions.supercomputing.org_SC19_sess130_ws_ipdrm107@linklings.com
SUMMARY:Characterizing the Performance of Executing Many-Tasks on Summit
DESCRIPTION:Workshop\n\nCharacterizing the Performance of Executing Many-T
 asks on Summit\n\nTurilli, Merzky, Naughton, Elwasif, Jha\n\nMany scientif
 ic workloads are comprised of many tasks, where each task is an independen
 t simulation or analysis of data. The execution of millions of tasks on he
 terogeneous HPC platforms requires scalable dynamic resource management an
 d multi-level scheduling. RADICAL-Pilot  (RP) -- an implementation of the 
 Pilot abstraction, addresses these challenges and serves as an effective r
 untime system to execute workloads comprised of many tasks. In this paper,
  we characterize the performance of executing many tasks using RP when int
 erfaced with JSM and PRRTE on Summit: RP is responsible for resource manag
 ement and task scheduling on acquired resource; JSM or PRRTE enact the pla
 cement of launching of scheduled tasks. Our experiments provide lower boun
 ds on the performance of RP when integrated with JSM and PRRTE. Specifical
 ly, for workloads comprised of homogeneous single-core, 15 minutes-long ta
 sks we find that: PRRTE scales better than JSM for > O(1000) tasks; PRRTE 
 overheads are negligible; and PRRTE supports optimizations that lower the 
 impact of overheads and enable resource utilization of 63% when executing 
 O(16K), 1-core tasks over 404 compute nodes.\n\nTag: Workshop Reg Pass, Co
 mpiler Analysis and Optimization, Middleware, Parallel Programming Languag
 es, Libraries, and Models, Runtime Systems\n\nRegistration Category: Works
 hop Reg Pass, Compiler Analysis and Optimization, Middleware, Parallel Pro
 gramming Languages, Libraries, and Models, Runtime Systems
URL:https://sc19.supercomputing.org/presentation/?id=ws_ipdrm107&sess=sess
 130
END:VEVENT
END:VCALENDAR