BEGIN:VCALENDAR VERSION:2.0 PRODID:-//wp-events-plugin.com//7.4.0.1//EN TZID:Asia/Kolkata X-WR-TIMEZONE:Asia/Kolkata BEGIN:VEVENT UID:224@cds.iisc.ac.in DTSTART;TZID=Asia/Kolkata:20260827T153000 DTEND;TZID=Asia/Kolkata:20260827T163000 DTSTAMP:20260826T154118Z 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-2/ SUMMARY:M.Tech Research Thesis Defense: CDS: "Asynchronous computing and lo w-precision approaches towards accelerating flow simulations" DESCRIPTION:DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES\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 approa ches towards accelerating flow simulations."\nThesis examiner : Prof. Ratn esh K. Shukla\, Mechanical Engineering\, IISc\nResearch Supervisor : Dr. K onduri Aditya\nDate &\; Time : August 27\, 2026\, 03:30 PM\nVenue : # 2 02 CDS Class room\n\n\n\nABSTRACT\n\nThe increasing computational cost of high-fidelity flow simulations at extreme scales has made communication ov erheads arising from data movement and synchronization a major bottleneck in modern high-performance computing. At the same time\, emerging GPU- and TPU-based architectures provide significantly higher throughput for low-p recision arithmetic compared to traditional double-precision computations. This thesis investigates asynchronous computing approaches and low-precis ion numerical frameworks towards accelerating compressible and reacting fl ow simulations on future exascale supercomputers.\n\nThe primary focus of this work is the development and evaluation of asynchronous numerical meth ods that relax communication and synchronization at a mathematical level w hile preserving the high-order accuracy of the underlying numerical scheme s. Previously 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. Tw o asynchronous algorithms are considered: one that avoids communication ov er a few predetermined time steps\, and another that initiates communicati on without enforcing synchronization. The numerical efficacy and scalabili ty of these 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 NACA001 2 airfoil. Scaling experiments performed on up to 18\,432 cores demonstrat e speed-ups of up to four times with respect to the baseline synchronous s olver while maintaining solution accuracy. These results demonstrate the a pplicability of AT schemes to established CFD solvers for improving scalab ility at extreme scales.\n\nThis work further extends the asynchronous com puting framework to discontinuous Galerkin (DG) methods for compressible r eacting flows. Although DG methods are attractive for their high arithmeti c intensity and their ability to accurately handle discontinuities such as shocks and detonations\, their scalability is also limited by communicati on bottlenecks arising from synchronization between processing elements (P Es). An asynchronous discontinuous Galerkin (ADG) method is developed for chemically reacting flows with detailed chemistry\, and new asynchrony-tol erant weighted essentially non-oscillatory (AT-WENO) limiters are proposed to accurately capture discontinuities in the presence of communication de lays near PE boundaries. The numerical properties of the ADG framework are evaluated for spontaneous ignition\, premixed flame propagation\, and det onation-wave propagation on a one-dimensional domain. The asynchronous sol ver accurately captures ignition fronts and discontinuities while incurrin g negligible numerical errors at PE boundaries. Preliminary scaling studie s further demonstrate the potential of the ADG method as a basis for highl y scalable DG-based solvers for massively parallel combustion simulations. \n\nIn addition to asynchronous algorithms\, this thesis also explores low -precision approaches for reacting-flow simulations motivated by the hardw are characteristics of modern accelerators. A low-precision framework is i nvestigated in which the chemical kinetics evaluations are performed in ha lf precision (FP16)\, while the nonlinear temperature solve is performed i n higher precision. The framework is assessed using lean hydrogen-air auto ignition with detailed chemical kinetics. Predictions of ignition delay an d the evolution of temperature\, heat-release rate\, and species mass frac tions show excellent agreement with FP64 reference solutions over a range of conditions. These preliminary findings demonstrate the feasibility of l ow-precision approaches for reacting-flow solvers while also identifying i mportant considerations regarding robustness and generalizability.\n\nOver all\, this thesis demonstrates that asynchronous computing methodologies a nd low-precision numerical approaches provide promising and complementary pathways towards improving the scalability and computational efficiency of next-generation flow solvers for exascale scientific computing.\n\n\n\nAL L ARE WELCOME CATEGORIES:Events,Thesis Defense END:VEVENT BEGIN:VTIMEZONE TZID:Asia/Kolkata X-LIC-LOCATION:Asia/Kolkata BEGIN:STANDARD DTSTART:20250827T153000 TZOFFSETFROM:+0530 TZOFFSETTO:+0530 TZNAME:IST END:STANDARD END:VTIMEZONE END:VCALENDAR