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:20260422T143138Z
LOCATION:B303
DTSTART;TZID=America/New_York:20241118T170000
DTEND;TZID=America/New_York:20241118T173000
UID:submissions.supercomputing.org_SC24_sess748_ws_pmbsf106@linklings.com
SUMMARY:Workload-Adaptive Scheduling for Efficient Use of Parallel File Sy
 stems in High-Performance Computing Clusters
DESCRIPTION:Alexander Goponenko (University of Central Florida), Benjamin 
 Allan and James Brandt (Sandia National Laboratories), and Damian Dechev (
 University of Central Florida)\n\nWhereas contentions within storage syste
 ms noticeably impact runtimes, shared bandwidth-type resources, such as Lu
 stre, pose challenges for high-performance computing cluster schedulers. A
 dditionally, accurately estimating job resource requirements, particularly
  related to I/O operations, remains a significant challenge for users. In 
 response to these challenges, we have developed a prototype that facilitat
 es I/O-aware scheduling in Slurm without imposing additional burdens on us
 ers. Accounting for the specific properties of this bandwidth-type resourc
 e, our system monitors real-time Lustre bandwidth utilization, estimates j
 ob I/O requirements, and dynamically adjusts to the demands placed on the 
 file system. Our workload-adaptive scheduler aims to maintain the bandwidt
 h utilization at a level that reflects the resource requirement of the job
  queue. We further enhance the efficacy of our approach by introducing a "
 two-group'' approximation technique that ensures efficient performance reg
 ardless of the availability of zero-throughput jobs. We demonstrate the ef
 fectiveness of our approach through evaluation on a real cluster.\n\nTag: 
 Accelerators, Modeling and Simulation, Performance Evaluation and/or Optim
 ization Tools\n\nRegistration Category: Workshop Reg Pass\n\nSession Chair
 s: Simon Hammond (National Nuclear Security Administration (NNSA)); Sascha
  Hunold (Technical University of Vienna); and Steven A. Wright (University
  of York, England)\n\n
END:VEVENT
END:VCALENDAR