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:193@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20260612T150000
DTEND;TZID=Asia/Kolkata:20260612T160000
DTSTAMP:20260507T141845Z
URL:https://cds.iisc.ac.in/events/ph-d-thesis-defense-102-cds-12-june-2026
 -scalable-real-space-finite-element-methods-and-algorithms-for-ab-initio-m
 aterial-modelling-in-the-exascale-era-applications-to-energy-materials/
SUMMARY:Ph.D: Thesis Defense: 102: CDS: 12\, June 2026 "Scalable real-space
  finite-element methods and algorithms for ab initio material modelling in
  the exascale era: Applications to energy materials"
DESCRIPTION:DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES\nPh.D. Thesis Def
 ense\n\n\n\nSpeaker: Mr. Kartick Ramakrishnan\nS.R. Number: 06-18-01-10-12
 -20-1-18495\nTitle: "Scalable real-space finite-element methods and algori
 thms for ab initio material modelling in\nthe exascale era: Applications t
 o energy materials"\nResearch Supervisor: Dr. Phani Motamarri\nDate &amp\;
  Time : June 12\, 2026 (Friday)\, 15:00 PM\nVenue : #102\, CDS Seminar Hal
 l\n\n\n\nABSTRACT\nOver the past six decades\, Kohn-Sham density functiona
 l theory (DFT) has transformed materials research by providing a quantum m
 echanical framework for accurate prediction of wide variety of ground-stat
 e properties. This predictive capability of DFT has made it ubiquitous acr
 oss chemistry\, condensed matter physics\, materials science\, and nanosci
 ence\, powering discoveries in catalysis\, energy conversion\, light-weigh
 t alloys\, electronic materials\, while consuming 30-40% of the world's hi
 gh performance computing resources today. Traditionally\, DFT has been emp
 loyed in high-throughput studies to establish structure–property relatio
 nships by screening large materials spaces\, typically involving material 
 systems with only a few hundred atoms. However\, despite rapid advances in
  high-performance computing\, state-of-the-art DFT methods are unable to f
 ully harness emerging exascale architectures\, thereby restricting simulat
 ions to length scales well below tens of nanometres required to capture co
 llective materials behaviour. Accessing these scales is crucial not only f
 or understanding complex phenomena in realistic materials systems but also
  for generating rich\, high-fidelity datasets that can drive machine-learn
 ing models towards predictive simulations across larger length and longer 
 time scales. This thesis seeks to bridge this gap by developing novel real
 -space computational methodologies and scalable algorithms that are inhere
 ntly suited to exascale architectures\, enabling accurate\, robust\, and m
 assively parallel ab initio simulations with a particular focus towards th
 eir applications to problems in energy storage and catalysis.\nTo address 
 these challenges\, this thesis first focuses on the development of a real-
 space finite-element-based methodology for density functional theory withi
 n the projector augmented-wave (PAW) formalism\, hereafter referred to as 
 PAW-FE. To the best of our knowledge\, this is the first real-space approa
 ch for DFT calculations\, combining the efficiency of PAW formalism involv
 ing smooth electronic fields with the ability of systematically improvable
  higher-order finite-element (FE) basis to achieve significant computation
 al gains across a wide-range of materials systems. Towards this\, we intro
 duce a local real-space formulation of the PAW energy functional that is n
 aturally amenable to finite-element discretisation resulting in a large-sp
 arse generalised eigenvalue problem (GHEP). A central computational challe
 nge in PAW-FE lies in efficiently computing the lowest N eigenpairs of thi
 s large\, sparse eigenproblem\, where N is proportional to the number of a
 toms and can exceed 50\,000 in large-scale simulations. The resulting eige
 nproblem is solved using a residual-based Chebyshev filtered subspace iter
 ation procedure (R-ChFSI)\, which is inherently tolerant to approximations
  in matrix-vector products. Leveraging this property\, we develop efficien
 t strategies that exploit the low-rank perturbation of the FE basis overla
 p matrix together with the reduced order quadrature rules to invert the di
 scretised PAW overlap matrix\, while utilizing the sparsity of both the lo
 cal and nonlocal parts of the resulting discretised matrices. Furthermore\
 , the robustness of R-ChFSI allows us to develop mixed precision strategie
 s to accelerate computation and communication\, combined with compute-comm
 unication overlap techniques that yield substantial performance gains on m
 odern GPU architectures without compromising accuracy. Together\, these ad
 vances enable scalable simulations of medium to large-scale materials syst
 ems. We further validate the accuracy and robustness of the proposed PAW-F
 E methodology against ABINIT a plane-wave DFT code\, demonstrating excelle
 nt agreement across representative materials systems.\n\nBuilding on these
  developments\, the thesis next presents the formulation of atomic forces 
 and cell stresses within the PAW-FE framework\, quantities that are essent
 ial for structure relaxation and ab initio molecular dynamics simulations.
  These correspond to the derivatives of the total energy functional with r
 espect to atomic positions and lattice vectors\, respectively. To this end
 \, we introduce a new energy functional that has the same stationary point
 s as the PAW energy functional and employ it to derive an expression for t
 he generalised force as a directional derivative of this energy functional
  with respect to a parameter that perturbs the underlying space. Consequen
 tly\, this formulation provides a unified expression for evaluating ionic 
 forces and unit-cell stresses\, while naturally incorporating Pulay contri
 butions. The resulting formulation is agnostic to the underlying discretis
 ation scheme employed\, making it broadly applicable across real-space met
 hods. We validate this approach within the PAW-FE formulation by benchmark
 ing against ABINIT\, demonstrating excellent agreement and confirming the 
 conservative nature of the computed forces and stresses.\n\nHaving establi
 shed the accuracy and robustness of our methodology for ground-state prope
 rties\, we proceed to assess the computational performance and scalability
  of the proposed methods relative to state-of-the-art (SOTA) DFT approache
 s. Towards this\, we conduct extensive CPU and GPU benchmarks against Quan
 tum Espresso (QE)\, a SOTA plane-wave open-source code\, and DFT-FE\, the 
 finite-element framework that was the workhorse behind the ACM Gordon Bell
  prize winning simulations. These benchmarks are performed on some of the 
 world’s foremost supercomputing platforms -- ALCF Aurora\, OLCF Frontier
 \, ALCF Polaris\, and NSM Param Pravega at IISc. We observe more than tenf
 old reduction in minimum wall time relative to QE for system sizes greater
  than 8\,000 electrons and more than sixfold reduction in computational co
 st compared to DFT-FE. These results highlight the ability of PAW-FE to ef
 ficiently exploit diverse exascale architectures\, exhibiting strong paral
 lel scalability and performance portability across heterogeneous computing
  environments. Furthermore\, the computational efficiency of our approach 
 enables high-fidelity simulations on modest\, in-house GPU clusters\, faci
 litating the rapid generation of rich datasets for machine-learning driven
  materials modelling.\n\nTo leverage the developed computational algorithm
 s for application problems\, this thesis next introduces a projected popul
 ation analysis methodology to extract quantitative chemical-bonding inform
 ation from large-scale real-space finite-element DFT calculations. Efficie
 nt computational strategies are devised to project the finite-element–di
 scretized Kohn–Sham orbitals onto atomic orbitals\, enabling the computa
 tion of projected overlap and Hamilton population that characterize bondin
 g interactions between atom pairs. This capability allows chemical-bonding
  analyses for systems containing thousands of atoms well beyond the reach 
 of traditional plane-wave approaches. We benchmark the accuracy and perfor
 mance of our implementation against LOBSTER\, a widely used population ana
 lysis code\, observing excellent agreement. Finally\, we demonstrate the p
 ractical utility of the framework by analysing H₂ chemisorption on silic
 on nanoclusters up to 10 nm in size and investigate the influence of carbo
 n alloying on the Si–H bond strength\, showing how carbon alloying alter
 s local bonding characteristics and influence the thermodynamics of hydrog
 en adsorption/desorption.\n\nTowards realizing the potential of the develo
 ped real-space computational methods for energy storage and catalysis appl
 ications\, the thesis next investigates strategies for accurately modellin
 g slabs and surfaces under external potential bias. Exploiting the capabil
 ity of real-space methods to accommodate generic boundary conditions\, a g
 eneralized framework for applying external bias across surfaces and interf
 aces within finite-element DFT is established. Two complementary strategie
 s are devised in this regard. First\, an external constant electric field 
 is introduced by modifying the DFT Hamiltonian with an auxiliary linear po
 tential\, while the electrostatic potential involved in the DFT problem is
  solved through a Poisson equation with zero-Neumann boundary conditions. 
 Second\, a desired potential bias is imposed directly by constraining the 
 electrostatic potential in a specified region\, thus allowing the direct s
 imulation of experimental conditions. We validate the constant-field imple
 mentation by comparing real-space finite-element DFT results with equivale
 nt plane-wave calculations on benchmark systems. Additionally we demonstra
 te comprehensive evaluation of the two strategies in terms of average grou
 nd-state properties such as surface and adsorption energies as a function 
 of external potential bias. Finally\, using this framework\, we present in
 itial simulations of lithium wetting at solid||electrolyte interfaces unde
 r applied bias demonstrating the influence of localized electric fields on
  interfacial energetics towards the development of design principles to op
 timize interfacial contact between lithium electrodes and solid electrolyt
 es.\nIn conclusion\, this thesis advances the boundaries of first-principl
 es materials modelling by developing a new generation of scalable\, real-s
 pace computational methods capable of harnessing the power of emerging exa
 scale architectures. The research introduces formulations and algorithms t
 hat are accurate\, efficient\, and scalable—enabling reliable simulation
 s of complex materials at unprecedented scales. Together\, these contribut
 ions lay the foundation for predictive\, exascale-ready DFT simulations th
 at bridge the gap between ab initio theory and real-world materials design
 .\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:20250612T150000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR