uu.seUppsala University Publications
Change search
ReferencesLink to record
Permanent link

Direct link
A GPU code for analytic continuation through a sampling method
Uppsala University, Disciplinary Domain of Science and Technology, Physics, Department of Physics and Astronomy. (materials theory)
Uppsala University, Disciplinary Domain of Science and Technology, Physics, Department of Physics and Astronomy. (materials theory)
Uppsala University, Disciplinary Domain of Science and Technology, Physics, Department of Physics and Astronomy. Radboud University. (Materials theory)
Uppsala University, Disciplinary Domain of Science and Technology, Physics, Department of Physics and Astronomy. (Materials theory)
2016 (English)In: SoftwareX, ISSN 2352-7110Article in journal (Refereed) In press
Abstract [en]

We here present a code for performing analytic continuation of fermionic Green’s functions and self-energies as well as bosonic susceptibilities on a graphics processing unit (GPU). The code is based on the sampling method introduced by Mishchenko et al. (2000), and is written for the widely used CUDA platform from NVidia. Detailed scaling tests are presented, for two different GPUs, in order to highlight the advantages of this code with respect to standard CPU computations. Finally, as an example of possible applications, we provide the analytic continuation of model Gaussian functions, as well as more realistic test cases from many-body physics.

Place, publisher, year, edition, pages
Elsevier, 2016.
Keyword [en]
GPU, Analytic continuation, Parallelization, Green’s function
National Category
Natural Sciences
Research subject
Physics
Identifiers
URN: urn:nbn:se:uu:diva-308687DOI: 10.1016/j.softx.2016.08.003OAI: oai:DiVA.org:uu-308687DiVA: diva2:1050638
Funder
eSSENCE - An eScience CollaborationKnut and Alice Wallenberg FoundationSwedish Research CouncilSwedish National Infrastructure for Computing (SNIC)
Available from: 2016-11-29 Created: 2016-11-29 Last updated: 2016-11-29
In thesis
1. Theoretical methods for the electronic structure and magnetism of strongly correlated materials
Open this publication in new window or tab >>Theoretical methods for the electronic structure and magnetism of strongly correlated materials
2017 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

In this work we study the interesting physics of the rare earths, and the microscopic state after ultrafast magnetization dynamics in iron. Moreover, this work covers the development, examination and application of several methods used in solid state physics. The first and the last part are related to strongly correlated electrons. The second part is related to the field of ultrafast magnetization dynamics.

In the first part we apply density functional theory plus dynamical mean field theory within the Hubbard I approximation to describe the interesting physics of the rare-earth metals. These elements are characterized by the localized nature of the 4f electrons and the itinerant character of the other valence electrons. We calculate a wide range of properties of the rare-earth metals and find a good correspondence with experimental data. We argue that this theory can be the basis of future investigations addressing rare-earth based materials in general.

In the second part of this thesis we develop a model, based on statistical arguments, to predict the microscopic state after ultrafast magnetization dynamics in iron. We predict that the microscopic state after ultrafast demagnetization is qualitatively different from the state after ultrafast increase of magnetization. This prediction is supported by previously published spectra obtained in magneto-optical experiments. Our model makes it possible to compare the measured data to results that are calculated from microscopic properties. We also investigate the relation between the magnetic asymmetry and the magnetization.

In the last part of this work we examine several methods of analytic continuation that are used in many-body physics to obtain physical quantities on real energies from either imaginary time or Matsubara frequency data. In particular, we improve the Padé approximant method of analytic continuation. We compare the reliability and performance of this and other methods for both one and two-particle Green's functions. We also investigate the advantages of implementing a method of analytic continuation based on stochastic sampling on a graphics processing unit (GPU).

Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 2017. 109 p.
Series
Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1651-6214 ; 1461
Keyword
dynamical mean field theory (DMFT), Hubbard I approximation, strongly correlated systems, rare earths, lanthanides, photoemission spectra, ultrafast magnetization dynamics, analytic continuation, Padé approximant method, two-particle Green's functions, linear muffin tin orbitals (LMTO), density functional theory (DFT), cerium, stacking fault energy.
National Category
Natural Sciences Physical Sciences
Research subject
Physics with spec. in Atomic, Molecular and Condensed Matter Physics
Identifiers
urn:nbn:se:uu:diva-308699 (URN)978-91-554-9770-5 (ISBN)
Public defence
2017-02-03, Ång/10132, Häggsalen, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, 09:15 (English)
Opponent
Supervisors
Available from: 2017-01-12 Created: 2016-11-29 Last updated: 2017-01-17

Open Access in DiVA

No full text

Other links

Publisher's full texthttp://www.sciencedirect.com/science/article/pii/S2352711016300243

Search in DiVA

By author/editor
Schött, JohanLocht, Inka L. M.Di Marco, Igor
By organisation
Department of Physics and Astronomy
Natural Sciences

Search outside of DiVA

GoogleGoogle Scholar

Altmetric score

Total: 81 hits
ReferencesLink to record
Permanent link

Direct link