Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
O: Fachverband Oberflächenphysik
O 48: Focussed session: Theory and computation of electronic structure: new frontiers IV (jointly with HL, DS)
O 48.6: Vortrag
Mittwoch, 16. März 2011, 16:30–16:45, TRE Phy
A conventional, massively parallel eigensolver for electronic structure theory1 — •V. Blum1, M. Scheffler1, R. Johanni2, H. Lederer2, Th. Auckenthaler3, Th. Huckle3, H.-J. Bungartz3, L. Krämer4, P. Willems4, B. Lang4, and V. Havu5 — 1Fritz Haber Institute, Berlin — 2RZ Garching — 3TU München — 4BU Wuppertal — 5Aalto University, Helsinki
We demonstrate a robust large-scale, massively parallel conventional eigensolver for first-principles theory of molecules and materials. Despite much research into O(N) methods, standard approaches (Kohn-Sham or Hartree-Fock theory and excited-state formalisms) must still rely on conventional but robust O(N3) solvers for many system classes, most notably metals. In particular, our eigensolver overcomes parallel scalability limitations where standard implementations of certain steps (reduction to tridiagonal form, solution of reduced tridiagonal eigenproblem) can be a serious bottleneck already for a few hundred CPUs. We demonstrate scalable implementations of these and all other steps of the full generalized eigenvalue problem. Our largest example is a production run with 1046 Pt (heavy-metal) atoms [1] with converged all-electron accuracy in the numeric atom-centered orbital code FHI-aims,[2] but the implementation is generic and should easily be portable to other codes. [1] P. Havu et al., Phys. Rev. B 82, 161418 (2010). [2] V. Blum et al., Comp. Phys. Comm. 180, 2175 (2009). 1 ELPA research consortium, funded by German Ministry of Research and Education (BMBF). http://elpa.rzg.mpg.de