BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.4.3//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:221@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20260821T140000
DTEND;TZID=Asia/Kolkata:20260821T150000
DTSTAMP:20260820T115409Z
URL:https://cds.iisc.ac.in/events/m-tech-research-thesis-defense-cds-async
 hronous-computing-and-low-precision-approaches-towards-accelerating-flow-s
 imulations/
SUMMARY:Postponed: M.Tech Research Thesis Defense: CDS: "Asynchronous compu
 ting and low-precision approaches towards accelerating flow simulations"
DESCRIPTION:Please note that Mr. Aswin Kumar's M.Tech Research Thesis Defen
 se\, scheduled for August 21 at 2:00 PM\, has been postponed. We regret an
 y inconvenience caused. The updated schedule with detailed timings and ven
 ue will be announced shortly.\nDEPARTMENT OF COMPUTATIONAL AND DATA SCIENC
 ES\nM.Tech Research Thesis Defense\n\n\n\nSpeaker : Mr. Aswin Kumar A\nS.R
 . Number : 06-18-01-10-22-24-1-24417\nTitle : "Asynchronous computing and 
 low-precision approaches towards accelerating flow simulations."\nThesis e
 xaminer : Prof. Ratnesh K. Shukla\, Mechanical Engineering\, IISc\nResearc
 h Supervisor: Dr. Konduri Aditya\nDate &amp\; Time : August 21\, 2026\, 02
 :00 PM\nVenue : # 102 CDS Seminar Hall\n\n\n\nABSTRACT\n\nThe increasing c
 omputational cost of high-fidelity flow simulations at extreme scales has 
 made communication overheads arising from data movement and synchronizatio
 n a major bottleneck in modern high-performance computing. At the same tim
 e\, emerging GPU- and TPU-based architectures provide significantly higher
  throughput for low-precision arithmetic compared to traditional double-pr
 ecision computations. This thesis investigates asynchronous computing appr
 oaches and low-precision numerical frameworks towards accelerating compres
 sible and reacting flow simulations on future exascale supercomputers.\n\n
 The primary focus of this work is the development and evaluation of asynch
 ronous numerical methods that relax communication and synchronization at a
  mathematical level while preserving the high-order accuracy of the underl
 ying numerical schemes. Previously developed asynchrony-tolerant (AT) sche
 mes are incorporated into the high-order compressible flow solver COMP-SQU
 ARE in a multi-block framework for practically relevant flow problems in c
 omplex geometries. Two asynchronous algorithms are considered: one that av
 oids communication over a few predetermined time steps\, and another that 
 initiates communication without enforcing synchronization. The numerical e
 fficacy and scalability of these asynchronous algorithms are demonstrated 
 for several benchmark problems\, including isentropic advection of a vorte
 x\, the Taylor-Green vortex\, and the highly sensitive case of transitiona
 l flow over a NACA0012 airfoil. Scaling experiments performed on up to 18\
 ,432 cores demonstrate speed-ups of up to four times with respect to the b
 aseline synchronous solver while maintaining solution accuracy. These resu
 lts demonstrate the applicability of AT schemes to established CFD solvers
  for improving scalability at extreme scales.\n\nThis work further extends
  the asynchronous computing framework to discontinuous Galerkin (DG) metho
 ds for compressible reacting flows. Although DG methods are attractive for
  their high arithmetic intensity and their ability to accurately handle di
 scontinuities such as shocks and detonations\, their scalability is also l
 imited by communication bottlenecks arising from synchronization between p
 rocessing elements (PEs). An asynchronous discontinuous Galerkin (ADG) met
 hod is developed for chemically reacting flows with detailed chemistry\, a
 nd new asynchrony-tolerant weighted essentially non-oscillatory (AT-WENO) 
 limiters are proposed to accurately capture discontinuities in the presenc
 e of communication delays near PE boundaries. The numerical properties of 
 the ADG framework are evaluated for spontaneous ignition\, premixed flame 
 propagation\, and detonation-wave propagation on a one-dimensional domain.
  The asynchronous solver accurately captures ignition fronts and discontin
 uities while incurring negligible numerical errors at PE boundaries. Preli
 minary scaling studies further demonstrate the potential of the ADG method
  as a basis for highly scalable DG-based solvers for massively parallel co
 mbustion simulations.\n\nIn addition to asynchronous algorithms\, this the
 sis also explores low-precision approaches for reacting-flow simulations m
 otivated by the hardware characteristics of modern accelerators. A low-pre
 cision framework is investigated in which the chemical kinetics evaluation
 s are performed in half precision (FP16)\, while the nonlinear temperature
  solve is performed in higher precision. The framework is assessed using l
 ean hydrogen-air autoignition with detailed chemical kinetics. Predictions
  of ignition delay and the evolution of temperature\, heat-release rate\, 
 and species mass fractions show excellent agreement with FP64 reference so
 lutions over a range of conditions. These preliminary findings demonstrate
  the feasibility of low-precision approaches for reacting-flow solvers whi
 le also identifying important considerations regarding robustness and gene
 ralizability.\n\nOverall\, this thesis demonstrates that asynchronous comp
 uting methodologies and low-precision numerical approaches provide promisi
 ng and complementary pathways towards improving the scalability and comput
 ational efficiency of next-generation flow solvers for exascale scientific
  computing.\n\n\n\nALL ARE WELCOME
CATEGORIES:Events,Thesis Defense
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20250821T140000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR