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:20230124T171525Z
LOCATION:C148
DTSTART;TZID=America/Chicago:20221113T134400
DTEND;TZID=America/Chicago:20221113T134800
UID:submissions.supercomputing.org_SC22_sess425_ws_whpc107@linklings.com
SUMMARY:Performance Optimizations within GPU Memory Constraints in Nyx Cos
 mology Code
DESCRIPTION:Workshop\n\nPerformance Optimizations within GPU Memory Constr
 aints in Nyx Cosmology Code\n\nSexton\n\nThe Nyx cosmology code is used to
  simulate the formation of large scale structure in Lyα simulations of the
  universe. As the dark matter particles begin clustering, the cost of a si
 ngle time step grows due to the load imbalance. These highly clustered reg
 ions can also prohibit fitting the entire problem in GPU HBM. If the entir
 e problem does not fit on the HBM, Nyx should use managed memory, and the 
 cost of each time step becomes dependent on the host-device memory bandwid
 th. This also imposes dynamic\nrestrictions on the best domain decompositi
 on for other physic components such as the heating-cooling and the hydrody
 namics solve.\n\nIn this talk, we will focus on different performance char
 acteristics of Nyx when constrained by load imbalance and the GPU memory c
 apacity, as well as the current approach in Nyx to optimizing this.\n\nSes
 sion Format: Recorded\n\nTag: Diversity Equity Inclusion (DEI), Education 
 and Training and Outreach\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR