Recorded 03 May 2022. Francois Gygi of University of California, Davis, Computer Science, presents "Generating Reference Data and Controlling Accuracy in DFT and Hybrid DFT Simulations" at IPAM's Large-Scale Certified Numerical Methods in Quantum Mechanics Workshop. Abstract: Density Functional Theory (DFT) simulations rely on several numerical approximations, notably the use of pseudopotentials that reduce the computational cost by removing core electrons, and the use of finite basis sets. We describe an all-electron approach used to produce reference results that can be used to validate pseudopotentials, and provide a systematic evaluation of the convergence of basis sets. We also address the issue of accelerating hybrid DFT simulations. Various acceleration schemes have been proposed to reduce computational cost using localization properties of orbitals, thus enabling hybrid-DFT first-principles molecular dynamics simulations. We discuss approaches that allow for error control when truncating localized orbitals. Learn more online at: http://www.ipam.ucla.edu/programs/workshops/workshop-iii-large-scale-certified-numerical-methods-in-quantum-mechanics/?tab=schedule