Towards Exascale Hybrid Electronic-Structure Theory Calculations beyond 10,000 Atoms

The advent of exascale computing paves the way for more accurate first-principles predictions that can target complex materials under realistic conditions. However, this requires electronic-structure theory codes that are specifically optimized to harvest the computational power of massively parallel CPU/GPU architectures. This challenge is addressed in the EU Centre of Excellence NOMAD by developing code-independent libraries for the computationally dominant operations. In this work, we present recent algorithmic advancements in memory parallelization and load-distribution that improve the performance of hybrid-functional calculations by two orders of magnitude compared to the original implementation in FHI-aims [1]. This linear-scaling evaluation of exact exchange integrals enables investigating systems with several 10,000 atoms, in the limit of which the cubically scaling eigenvalue solvers become computationally dominant. In this regard, we report recent optimizations of the ELPA library [2], including their adaption to several different GPU architectures, that help in mitigating this hurdle.

[1] S. V. Levchenko et al., Comp. Phys. Comm. 192, 60 (2015).

[2] P. Kus et al., Par. Comp. 85, 167 (2019).