BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20220812T074335Z
LOCATION:Samarkand Room
DTSTART;TZID=Europe/Stockholm:20220629T113000
DTEND;TZID=Europe/Stockholm:20220629T120000
UID:submissions.pasc-conference.org_PASC22_sess157_msa120@linklings.com
SUMMARY:A GPU-Based, Industrial-Grade Compositional Reservoir Simulator
DESCRIPTION:Minisymposium\n\nA GPU-Based, Industrial-Grade Compositional R
 eservoir Simulator\n\nFerrari\n\nRecently, graphics processing units (GPUs
 ) have been demonstrated to provide a significant performance benefit for 
 industrial applications where the computational bottleneck is the numerica
 l solution of partial differential equations. In this work we show how GPU
 s can be effectively used for complex problems, namely compositional simul
 ations needed in the energy industry to plan large investments for reservo
 ir development and CO2 storage. Solving compositional problems needs a com
 bination of efficient and scalable linear system accelerators and precondi
 tioner with effective phase equilibrium solver based on cubic equation of 
 state. First, we describe the motivations for the selected nonlinear formu
 lation, including the choice of primary variables and iteration scheme. Fo
 llowing industry standards, the implementation supports both fully implici
 t methods and adaptive implicit methods. We then present performance resul
 ts on an example sector model and simplified synthetic case designed to al
 low a detailed examination of runtime and memory scaling with respect to t
 he number of hydrocarbon components and model size, as well as the number 
 of processors. We finally show results from two complex asset models (synt
 hetic and real) and examine performance scaling with respect to GPU genera
 tion, demonstrating that performance correlates strongly with GPU memory b
 andwidth.\n\nDomain: Computer Science and Applied Mathematics, Engineering
 , Physics
END:VEVENT
END:VCALENDAR