BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:195@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20260518T093000
DTEND;TZID=Asia/Kolkata:20260518T103000
DTSTAMP:20260511T131620Z
URL:https://cds.iisc.ac.in/events/m-tech-research-thesis-colloquium-cds-as
 ynchronous-computing-and-low-precision-approaches-towards-accelerating-flo
 w-simulations/
SUMMARY:M.Tech Research Thesis {Colloquium}: CDS: "Asynchronous computing a
 nd low-precision approaches towards accelerating flow simulations."
DESCRIPTION:\n\nSpeaker : Mr. Aswin Kumar A\nS.R. Number : 06-18-01-10-22-2
 4-1-24417\nTitle : "Asynchronous computing and low-precision approaches to
 wards accelerating flow simulations."\nResearch Supervisor : Dr. Konduri A
 ditya\nDate &amp\; Time : May 18\, 2026\, 09.30 AM\nVenue : # 102 CDS Semi
 nar Hall\n\n\n\nABSTRACT\n\nThe increasing computational cost of high-fide
 lity flow simulations at extreme scales has made communication overheads a
 rising from data movement and synchronization a major bottleneck in modern
  high-performance computing. At the same time\, emerging GPU- and TPU-base
 d architectures provide significantly higher throughput for low-precision 
 arithmetic compared to traditional double-precision computations. This the
 sis investigates asynchronous computing approaches and low-precision numer
 ical frameworks towards accelerating compressible and reacting flow simula
 tions on future exascale supercomputers.\n\nThe primary focus of this work
  is the development and evaluation of asynchronous numerical methods that 
 relax communication and synchronization at a mathematical level while pres
 erving the high-order accuracy of the underlying numerical schemes. Previo
 usly developed asynchrony-tolerant (AT) schemes are incorporated into the 
 high-order compressible flow solver COMP-SQUARE in a multi-block framework
  for practically relevant flow problems in complex geometries. Two asynchr
 onous algorithms are considered: one that avoids communication over a few 
 predetermined time steps\, and another that initiates communication withou
 t enforcing synchronization. The numerical efficacy and scalability of the
 se asynchronous algorithms are demonstrated for several benchmark problems
 \, including isentropic advection of a vortex\, the Taylor-Green vortex\, 
 and the highly sensitive case of transitional flow over a NACA0012 airfoil
 . Scaling experiments performed on up to 18\,432 cores demonstrate speed-u
 ps of up to four times with respect to the baseline synchronous solver whi
 le maintaining solution accuracy. These results demonstrate the applicabil
 ity of AT schemes to established CFD solvers for improving scalability at 
 extreme scales.\n\nThis work further extends the asynchronous computing fr
 amework to discontinuous Galerkin (DG) methods for compressible reacting f
 lows. Although DG methods are attractive for their high arithmetic intensi
 ty and their ability to accurately handle discontinuities such as shocks a
 nd detonations\, their scalability is also limited by communication bottle
 necks arising from synchronization between processing elements (PEs). An a
 synchronous discontinuous Galerkin (ADG) method is developed for chemicall
 y reacting flows with detailed chemistry\, and new asynchrony-tolerant wei
 ghted essentially non-oscillatory (AT-WENO) limiters are proposed to accur
 ately capture discontinuities in the presence of communication delays near
  PE boundaries. The numerical properties of the ADG framework are evaluate
 d for spontaneous ignition\, premixed flame propagation\, and detonation-w
 ave propagation on a one-dimensional domain. The asynchronous solver accur
 ately captures ignition fronts and discontinuities while incurring negligi
 ble numerical errors at PE boundaries. Preliminary scaling studies further
  demonstrate the potential of the ADG method as a basis for highly scalabl
 e DG-based solvers for massively parallel combustion simulations.\n\nIn ad
 dition to asynchronous algorithms\, this thesis also explores low-precisio
 n approaches for reacting-flow simulations motivated by the hardware chara
 cteristics of modern accelerators. A low-precision framework is investigat
 ed in which the chemical kinetics evaluations are performed in half precis
 ion (FP16)\, while the nonlinear temperature solve is performed in higher 
 precision. The framework is assessed using lean hydrogen-air autoignition 
 with detailed chemical kinetics. Predictions of ignition delay and the evo
 lution of temperature\, heat-release rate\, and species mass fractions sho
 w excellent agreement with FP64 reference solutions over a range of condit
 ions. These preliminary findings demonstrate the feasibility of low-precis
 ion approaches for reacting-flow solvers while also identifying important 
 considerations regarding robustness and generalizability.\n\nOverall\, thi
 s thesis demonstrates that asynchronous computing methodologies and low-pr
 ecision numerical approaches provide promising and complementary pathways 
 towards improving the scalability and computational efficiency of next-gen
 eration flow solvers for exascale scientific computing.\n\n\n\nALL ARE WEL
 COME
CATEGORIES:Events,MTech Research Thesis Colloquium
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20250518T093000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR