BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143140Z
LOCATION:B312
DTSTART;TZID=America/New_York:20241117T143000
DTEND;TZID=America/New_York:20241117T150000
UID:submissions.supercomputing.org_SC24_sess739_ws_ss116@linklings.com
SUMMARY:Vendor-neutral and Production-grade Job Power Management in High P
 erformance Computing
DESCRIPTION:Naman Kulshresthha (Clemson University); Tapasya Patki, Jim Ga
 rlick, and Mark Grondona (Lawrence Livermore National Laboratory (LLNL)); 
 and Rong Ge (Clemson University)\n\nPower management and energy efficiency
  are critical\nresearch areas for exascale computing and beyond, necessita
 ting\nreliable telemetry and control for distributed systems. Despite this
 \nneed, existing approaches present several limitations precluding\ntheir 
 adoption in production. These limitations include, but are\nnot limited to
 , lack of portability due to vendor-specific and\nclosed-source solutions,
  lack of support for non-MPI applications,\nand lack of user-level customi
 zation.\nWe present a job-level power management framework based\non Flux.
  We introduce flux-power-monitor and demonstrate\nits effectiveness on the
  Lassen (IBM Power AC922) and Tioga\n(HPE Cray EX235A) systems with a low 
 average overhead\nof 0.4%. We also present flux-power-manager, where we\nd
 iscuss a proportional sharing policy and introduce a hierarchical\nFFT-bas
 ed dynamic power management algorithm (FPP). We\ndemonstrate that FPP redu
 ces energy by 1% compared to\nproportional sharing, and by 20% compared to
  the default IBM\nstatic power capping policy.\n\nTag: Energy Efficiency, 
 HPC Infrastructure, Sustainability\n\nRegistration Category: Workshop Reg 
 Pass\n\nSession Chairs: Cate Berard (US Department of Energy); James H. Ro
 gers (Oak Ridge National Laboratory (ORNL)); Fumiyoshi Shoji (RIKEN Center
  for Computational Science (R-CCS), Center for Computational Science); Mic
 hèle Weiland (EPCC, The University of Edinburgh; The University of Edinbur
 gh); and Mike Woodacre (Hewlett Packard Enterprise (HPE))\n\n
END:VEVENT
END:VCALENDAR
