Solvers for $\mathcal{O} (N)$ Electronic Structure in the Strong Scaling Limit
This work addresses the challenge of strong scaling in electronic structure calculations for materials science, enabling efficient use of massive parallelism for small systems.
The authors present a hybrid OpenMP/Charm++ framework that achieves strong scaling for O(N) electronic structure calculations, scaling to hundreds of cores per molecule for small water clusters (e.g., P/N ≈ 819 for 30 molecules).
We present a hybrid OpenMP/Charm++ framework for solving the $\mathcal{O} (N)$ Self-Consistent-Field eigenvalue problem with parallelism in the strong scaling regime, $P\gg{N}$, where $P$ is the number of cores, and $N$ a measure of system size, i.e. the number of matrix rows/columns, basis functions, atoms, molecules, etc. This result is achieved with a nested approach to Spectral Projection and the Sparse Approximate Matrix Multiply [Bock and Challacombe, SIAM J.~Sci.~Comput. 35 C72, 2013], and involves a recursive, task-parallel algorithm, often employed by generalized $N$-Body solvers, to occlusion and culling of negligible products in the case of matrices with decay. Employing classic technologies associated with generalized $N$-Body solvers, including over-decomposition, recursive task parallelism, orderings that preserve locality, and persistence-based load balancing, we obtain scaling beyond hundreds of cores per molecule for small water clusters ([H${}_2$O]${}_N$, $N \in \{ 30, 90, 150 \}$, $P/N \approx \{ 819, 273, 164 \}$) and find support for an increasingly strong scalability with increasing system size $N$.