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:20200129T163557Z
LOCATION:603
DTSTART;TZID=America/Denver:20191118T110000
DTEND;TZID=America/Denver:20191118T113000
UID:submissions.supercomputing.org_SC19_sess122_ws_pmbsf114@linklings.com
SUMMARY:Enhancing Monte Carlo Proxy Applications on GPUs
DESCRIPTION:Workshop\n\nEnhancing Monte Carlo Proxy Applications on GPUs\n
 \nShriver, Lee, Hamilton, Vetter, Watson\n\nIn Monte Carlo neutron transpo
 rt simulations, a computational routine commonly known as the "cross-secti
 on lookup" has been identified as being the most computationally expensive
  part of these applications. A tool which is commonly used as a proxy appl
 ication for these routines, named "XSBench", was created to simulate popul
 ar algorithms used in these routines on CPUs. Currently, however, as GPU-b
 ased HPC resources have become more widely available, there has been signi
 ficant interest and efforts invested in moving these traditionally CPU-bas
 ed simulations to GPUs. Unfortunately, the algorithms commonly used in the
  cross-section lookup routine were originally devised and developed for CP
 U-based platforms, and have seen limited study on GPUs to date. Additional
 ly, platforms such as XSBench implement approximations which may have a ne
 gligible effect on CPUs, but may be quite impactful to performance on GPUs
  given the more resource-limited nature of the latter. As a result, we hav
 e created VEXS, a new tool for modeling the cross-section lookup routine w
 hich removes or at least reduces the approximations made by XSBench in ord
 er to provide a more realistic prediction of algorithm performance on GPUs
 . In this paper, we detail our efforts to remove and reduce these approxim
 ations, show the resulting improvement in performance prediction in compar
 ison to a reference production code, Shift, and provide some basic profili
 ng analysis of the resulting application.\n\nTag: Workshop Reg Pass, Bench
 marks, Performance, Scientific Computing, Simulation\n\nRegistration Categ
 ory: Workshop Reg Pass, Benchmarks, Performance, Scientific Computing, Sim
 ulation
URL:https://sc19.supercomputing.org/presentation/?id=ws_pmbsf114&sess=sess
 122
END:VEVENT
END:VCALENDAR