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:20260202T201258Z
LOCATION:Second Floor Atrium
DTSTART;TZID=America/Chicago:20251121T080000
DTEND;TZID=America/Chicago:20251121T120000
UID:submissions.supercomputing.org_SC25_sess620_post159@linklings.com
SUMMARY:Enabling Real-Time, Extreme-Scale Bayesian Inference: FFT-Based GP
 U-Accelerated Matrix-Vector Products for Block-Triangular Toeplitz Matrice
 s
DESCRIPTION:Sreeram Venkat and Omar Ghattas (The University of Texas at Au
 stin)\n\nAdjoint-based, matrix-free Newton-Krylov methods have long been t
 he gold standard for solving high-dimensional, ill-posed inverse problems.
  These methods require a pair of forward and adjoint PDE solves per iterat
 ion, usually making them intractable for real-time inference and predictio
 n. We present FFTMatvec, an FFT-based GPU-accelerated algorithm that explo
 its intrinsic problem structure to enable real-time, high-fidelity, extrem
 e-scale inference and prediction for linear autonomous dynamical systems. 
 This algorithm was used to solve a Bayesian inverse problem for tsunami ea
 rly warning with over one billion parameters in under 0.2 seconds. The app
 lication is performance-portable and open-source; scaling results are pres
 ented for up to 4,096 GPUs on OLCF's Frontier and NERSC's Perlmutter super
 computers. On 512 GPUs, FFTMatvec achieves more than a 200,000x speedup ov
 er state-of-the-art matrix-free adjoint-based methods. Communication-aware
  partitioning and dynamic mixed precision provide additional performance b
 oosts. Other application areas include nuclear treaty verification and mon
 itoring atmospheric CO2.\n\nTag: Research & ACM SRC Posters\n\nRegistratio
 n Category: Technical Program Reg Pass\n\nSession Chairs: Kento Sato (RIKE
 N Center for Computational Science (R-CCS)); Anja Gerbes (Georg-August-Uni
 versität Göttingen); and Chris Schlipalius (Pawsey Supercomputing Research
  Centre; Commonwealth Scientific and Industrial Research Organisation (CSI
 RO), Australia)\n\n
END:VEVENT
END:VCALENDAR