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:20241120T082410Z
LOCATION:HG F 1
DTSTART;TZID=Europe/Stockholm:20240603T173000
DTEND;TZID=Europe/Stockholm:20240603T180000
UID:submissions.pasc-conference.org_PASC24_sess173_pap133@linklings.com
SUMMARY:Leveraging the High Bandwidth of Last-Level Cache for HPC Seismic 
 Imaging Applications
DESCRIPTION:Paper\n\nPavel Plotnitskii (King Abdullah University of Scienc
 e and Technology); Louis Beaurepaire (Polytech Lyon); Long Qu (China Telec
 om Cloud Technology Corporation Limited); and Kadir Akbudak, Hatem Ltaief,
  and David Keyes (King Abdullah University of Science and Technology)\n\nW
 e solve the 3D acoustic wave equation using the finite-difference time-dom
 ain (FDTD) formulation in both first and second order. The FDTD approach i
 s expressed as a stencil-based computational scheme with a long-range disc
 retization, i.e., 8th order in space and 2nd order in time, which is routi
 nely used in the oil and gas industry and environmental geophysics for hig
 h subsurface imaging fidelity purposes. Absorbing Boundary Conditions (ABC
 s) are employed to attenuate reflections from artificial boundaries. The h
 igh order discretization engenders extensive data movement across the memo
 ry subsystem and may consequently impact the kernel throughput due to the 
 inherent memory-bound behavior of the stencil operator, especially on syst
 ems facing memory starvation. The first-order formulation of the 3D acoust
 ic equation further exacerbates this phenomenon because it calculates both
  the pressure and velocity fields, which corresponds to 1.6X the memory fo
 otprint of second-order formulation. To address this memory bottleneck, we
  design, implement, and deploy the multicore wavefront diamond tiling with
  temporal blocking (MWD-TB) to boost the performance of seismic wavefield 
 modeling by exploiting spatial&temporal data reuse. MWD-TB leverages the l
 arge capacity of last-level cache (LLC) of modern x86 systems and extracts
  high bandwidth memory from the underlying architecture. We demonstrate th
 e numerical accuracy of MWD-TB on the Salt3D model from the Society of Exp
 loration Geophysicists. Our MWD-TB implementations for the first- and seco
 nd-order FDTD formulations achieve speedups of up to 3.5X and 3X on a larg
 e grid size on AMD systems equipped with large LLC, respectively, compared
  to the traditional spatial blocking method alone.\n\nDomain: Climate, Wea
 ther, and Earth Sciences\n\nSession Chair: Thorsten Kurth (NVIDIA Inc.)
END:VEVENT
END:VCALENDAR