October-2024 online developer meeting

When:

October 8 and 9 afternoons (Danish time)

Announcement:

https://listserv.fysik.dtu.dk/pipermail/gpaw-users/2024-September/007312.html

Note

Talks are approximately 15 + 10 minutes. Please try to focus your talks on the equations and algorithms you need to solve and the challenges implementing those in the code.

Program

Tuesday

14:00:

Jens Jørgen Mortensen

Modernizing the code-base (slides)

Discussion of changes taking place in the code at the moment and how that will affect all of us in the future.

14:25:

Aleksei Ivanov

UPAW method, wavefunction theories and quantum computing prospects

In this presentation, I will introduce the unitary projector augmented-wave method (UPAW) and its generalization on many-body wave functions. Then I will discuss the benefits UPAW brings in the development of the electronic structure algorithms with focus on many-body theories and quantum computation. I will show a few examples on how one can calculate the MP2 energy with GPAW and the interface with other electronic structure software such as PySCF to carry out more complicated wavefunction calculations using, for example, coupled cluster theories. I will finish the talk with prospects on using GPAW for quantum embedding theories such as the density matrix embedding theory.

14:50:

Marko Melander

Constant inner potential (CIP) DFT method (extension of SJM)

CIP-DFT and its further development as well as working towards integrating Poisson-Boltzmann models in the main GPAW version.

15:15:

Break

15:25:

Colin Baker

Porting SJM into planewave mode

15:50:

Kyle Bystrom

PAW Implementation of Nonlocal, Machine-Learned Density Functionals

Description: I will discuss an approach for efficiently evaluating nonlocal density functionals within the PAW formalism and the application of this approach to machine learning-based functionals. I will also mention work in progress to further optimize this algorithm and plans to extend it to van der Waals functionals and orbital-dependent features.

16:15:

Pooria Dabbaghi

Implementing an inverse Kohn-Sham scheme within the PAW formalism

16:40:

Anubhab Haldar

Improved basis sets for LCAO

Wednesday

14:00:

Mikael Kuisma

Running calculations on GPUs

14:25:

Tuomas Rossi

Building GPAW for AMD and NVIDIA GPUs

I will discuss the portability approach in GPAW to support both NVIDIA and AMD GPUs.

14:50:

Ask Hjorth Larsen

Testing the code on our new CI-runner + response-code refactoring

15:15:

Break

15:25:

Gianluca Levi

Direct optimization methods for excited state calculations

The current status of the implementation, what is missing and what are the new features that I’d like to implement next.

15:50:

Gianluca Levi

Self-interaction correction

16:15:

Vladimír Zobač

Ehrenfest molecular dynamics with LCAO basis

16:40:

General discussion

17:05:

The end