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:20260522T150116Z
LOCATION:C140/142
DTSTART;TZID=America/Chicago:20181114T133000
DTEND;TZID=America/Chicago:20181114T140000
UID:submissions.supercomputing.org_SC18_sess204_pap502@linklings.com
SUMMARY:Simulating the Wenchuan Earthquake with Accurate Surface Topograph
 y on Sunway TaihuLight
DESCRIPTION:Bingwei Chen, Haohuan Fu, Yanwen Wei, and Conghui He (Tsinghua
  University; National Supercomputing Center, Wuxi); Wenqiang Zhang (Univer
 sity of Science and Technology of China); Yuxuan Li (Tsinghua University; 
 National Supercomputing Center, Wuxi); Wubin Wan and Wei Zhang (National S
 upercomputing Center, Wuxi); Lin Gan (Tsinghua University; National Superc
 omputing Center, Wuxi); Wei Zhang and Zhenguo Zhang (Southern University o
 f Science and Technology, China); Guangwen Yang (Tsinghua University; Nati
 onal Supercomputing Center, Wuxi); and Xiaofei Chen (Southern University o
 f Science and Technology, China)\n\nThis paper reports our efforts on perf
 orming 50-m resolution earthquake simulation of the Wenchuan Earthquake (M
 s 8.0, China) on Sunway TaihuLight. To accurately capture the surface topo
 graphy, we adopt a curvilinear grid finite-difference method with a tracti
 on image free surface implementation and redesign the algorithm to reduce 
 memory access costs for heterogeneous many-core architectures. We then der
 ive a performance model of our algorithm to guide and drive the further op
 timization and tuning of various parameters using a genetic algorithm. A d
 ata layout transformation is also proposed to improve the direct memory ac
 cess (DMA) efficiency further. Our efforts improve the simulation efficien
 cy from 0.05% to 7.6%, with a sustained performance of 9.07 Pflops using t
 he entire machine of the Sunway TaihuLight (over 10 million cores), and a 
 large-scale simulation of the Wenchuan earthquake with accurate surface to
 pography and improved coda wave effects.\n\nTag: Algorithms, Applications,
  Computational Physics, Scientific Computing\n\nRegistration Category: Tec
 h Program Reg Pass\n\nSession Chair: Erik Draeger (Lawrence Livermore Nati
 onal Laboratory (LLNL), Center for Applied Scientific Computing)\n\n
END:VEVENT
END:VCALENDAR
