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:20260422T000711Z
LOCATION:301-302-303
DTSTART;TZID=America/Denver:20231116T163000
DTEND;TZID=America/Denver:20231116T170000
UID:submissions.supercomputing.org_SC23_sess162_pap145@linklings.com
SUMMARY:Enabling Real World Scale Structural Superlubricity All-Atom Simul
 ation on the Next-Generation Sunway Supercomputer
DESCRIPTION:Xiaohui Duan (Shandong University, China; National Supercomput
 ing Center in Wuxi); Jin Wang (International School for Advanced Studies, 
 Italy); Ping Gao (Tsinghua University, National Supercomputing Center in W
 uxi); Ming Ma (Tsinghua University); Lin Gan (Tsinghua University, Nationa
 l Supercomputing Center in Wuxi); Xin Liu (National Supercomputing Center 
 in Wuxi); Haohuan Fu and Wei Xue (Tsinghua University, National Supercompu
 ting Center in Wuxi); Dexun Chen (National Supercomputing Center in Wuxi);
  Guangwen Yang (Tsinghua University, National Supercomputing Center in Wux
 i); and Weiguo Liu (Shandong University, China; National Supercomputing Ce
 nter in Wuxi)\n\nMolecular dynamics (MD) simulation can provide an afforda
 ble way for inspecting microscopic phenomena, which is a powerful compleme
 nt to real-world experiments. But the spatial scale of MD simulations is u
 sually magnitudes smaller than experiment systems. In this paper, we prese
 nt our work, redesigning the widely used inter-layer potential in structur
 al superlubricity. By carrying out a specialized neighbor list for inter-l
 ayer potential computation, the total memory access amount is reduced sign
 ificantly. Besides, a simple but efficient vectorization strategy is imple
 mented based on the new neighbor list. In the extreme case, our work can s
 cale to 38 million cores to achieve a sustainable performance of 61 PFLOPS
 , enabling a simulation of a superlubricity system of 32 um^2 with 7.2 bil
 lion atoms at 4.75 ns/day, which is 11,834 times of reported largest scale
  simulation in superlubricity systems in contact area and almost ten times
  faster in time-to-solution.\n\nTag: Accelerators, Applications, Architect
 ure and Networks, Modeling and Simulation\n\nRegistration Category: Tech P
 rogram Reg Pass\n\nReproducibility Badges: Artifact Available\n\nSession C
 hair: Sudip Seal (Oak Ridge National Laboratory (ORNL))\n\n
END:VEVENT
END:VCALENDAR
