arXiv subjects

Bernd Hartke

Publications and source records attributed to Bernd Hartke.

5 recordsLinked to original sources

arXiv2019

Adaptive parallelism with RMI: Idle high-performance computing resources can be completely avoided

In practice, standard scheduling of parallel computing jobs almost always leaves significant portions of the available hardware unused, even with many jobs still waiting in the queue. The simple reason is that the resource requests of these waiting jobs are fixed and do not match the available, unused resources. However, with alternative but existing and well-established techniques it is possible to achieve a fully automated, adaptive parallelism that does not need pre-set, fixed resources. Here, we demonstrate that such an adaptively parallel program can indeed fill in all such scheduling gaps, even in real-life situations on large supercomputers.

cs.DC↗

arXiv2018

Resonance dynamics of DCO ($\widetilde{X}\,{}^2A'$) simulated with the dynamically pruned discrete variable representation (DP-DVR)

Selected resonance states of the deuterated formyl radical in the electronic ground state ($\widetilde{X}\,{}^2A'$) are computed using our recently introduced dynamically pruned discrete variable representation (DP-DVR) [H. R. Larsson, B. Hartke and D. J. Tannor, J. Chem. Phys., 145, 204108 (2016)]. Their decay and asymptotic distributions are analyzed and, for selected resonances, compared to experimental results obtained by a combination of stimulated emission pumping (SEP) and velocity-map imaging of the product D atoms. The theoretical results show good agreement with the experimental kinetic energy distributions. The intramolecular vibrational energy redistribution (IVR) is analyzed and compared with previous results from an effective polyad Hamiltonian. Specifically, we analyzed the part of the wavefunction that remains in the interaction region during the decay. The results from the polyad Hamiltonian could mainly be confirmed. The C=O stretch quantum number is typically conserved, while the D-C=O bend quantum number decreases. Differences are due to strong anharmonic coupling such that all resonances have major contributions from several zero-order states. For some of the resonances, the coupling is so strong that no further zero-order states appear during the dynamics in the interaction region, even after propagating for 300 ps.

physics.chem-ph↗

arXiv2016

Efficient molecular quantum dynamics in coordinate and phase space using pruned bases

We present an efficient implementation of dynamically pruned quantum dynamics, both in coordinate space and in phase space. We combine the ideas behind the biorthogonal von Neumann basis (PvB) with the orthogonalized momentum-symmetrized Gaussians (Weylets) to create a new basis, projected Weylets, that takes the best from both methods. We benchmark pruned dynamics using phase-space-localized PvB, projected Weylets, and coordinate-space-localized DVR bases, with real-world examples in up to six dimensions. We show that coordinate-space localization is most important for efficient pruning and that pruned dynamics is much faster compared to unpruned, exact dynamics. Phase-space localization is useful for more demanding dynamics where many basis functions are required. There, projected Weylets offer a more compact representation than pruned DVR bases.

physics.chem-ph↗

arXiv2012

Empirical review of standard benchmark functions using evolutionary global optimization

We have employed a recent implementation of genetic algorithms to study a range of standard benchmark functions for global optimization. It turns out that some of them are not very useful as challenging test functions, since they neither allow for a discrimination between different variants of genetic operators nor exhibit a dimensionality scaling resembling that of real-world problems, for example that of global structure optimization of atomic and molecular clusters. The latter properties seem to be simulated better by two other types of benchmark functions. One type is designed to be deceptive, exemplified here by Lunacek's function. The other type offers additional advantages of markedly increased complexity and of broad tunability in search space characteristics. For the latter type, we use an implementation based on randomly distributed Gaussians. We advocate the use of the latter types of test functions for algorithm development and benchmarking.

cs.NE↗

arXiv1995

Global geometry optimization of clusters using a growth strategy optimized by a genetic algorithm

A new strategy for global geometry optimization of clusters is presented. Important features are a restriction of search space to favorable nearest-neighbor distance ranges, a suitable cluster growth representation with diminished correlations, and easy transferability of the results to larger clusters. The strengths and possible limitations of the method are demonstrated for Si10 using an empirical potential.

chem-ph↗