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:20181221T160903Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181115T083000
DTEND;TZID=America/Chicago:20181115T170000
UID:submissions.supercomputing.org_SC18_sess324_post197@linklings.com
SUMMARY:Multi-Client DeepIO for Large-Scale Deep Learning on HPC Systems
DESCRIPTION:Poster\nTech Program Reg Pass, Exhibits Reg Pass\n\nMulti-Clie
 nt DeepIO for Large-Scale Deep Learning on HPC Systems\n\nZhu, Chowdhury, 
 Fu, Moody, Mohror...\n\nWith the growth of computation power, leadership H
 igh-Performance Computing (HPC) systems can train larger datasets for Deep
  neural networks (DNNs) more efficiently. On HPC systems, a training datas
 et is on a parallel file system or node-local storage devices. However, no
 t all HPC clusters have node-local storage, and large mini-batch sizes str
 ess the read performance of parallel systems since the large datasets cann
 ot fit in file system caches. Thus, it is a challenge for training DNNs wi
 th large datasets on HPC systems.\n\nIn prior work, we proposed DeepIO to 
 mitigate the I/O pressure. DeepIO is designed to assist the mini-batch gen
 eration of TensorFlow. However, DeepIO does not support multiple training 
 workers on a single compute node. We address this gap with modification on
  DeepIO framework, and evaluate multi-client DeepIO performance against st
 ate-of-the-art in-memory file systems, compare DeepIO and TensorFlow data 
 loading API, and explore the potential of DeepIO in DNN training.
URL:https://sc18.supercomputing.org/presentation/?id=post197&sess=sess324
END:VEVENT
END:VCALENDAR