GPU Optimization of Large-Scale Eigenvalue Solver

  • Pavel KůsEmail author
  • Hermann Lederer
  • Andreas Marek
Conference paper
Part of the Lecture Notes in Computational Science and Engineering book series (LNCSE, volume 126)


We present a GPU implementation of a large-scale eigenvalue solver as a part of the ELPA library. We describe the methodology of utilizing the GPU accelerators within an already well optimized MPI-based code. We present numerical results using two different HPC systems equipped with modern GPU accelerators and show the performance benefits of the GPU version.



Part of this work is co-funded by BMBF grant 01IH15001 of the German Government.


  1. 1.
    T. Auckenthaler, V. Blum, H.-J. Bungartz, T. Huckle, R. Johanni, L. Krmer, B. Lang, H. Lederer, P.R. Willems, Parallel solution of partial symmetric eigenvalue problems from electronic structure calculations. Parallel Comput. 37, 783–794 (2011)CrossRefGoogle Scholar
  2. 2.
    A. Marek, V. Blum, R. Johanni, V. Havu, B. Lang, T. Auckenthaler, A. Heinecke, H.-J. Bungartz, H. Lederer, The ELPA library - scalable parallel eigenvalue solutions for electronic structure theory and computational science. J. Phys. Condens. Matter 26, 213201 (2014)CrossRefGoogle Scholar
  3. 3.
  4. 4.
    G.H. Golub, C.F.V. Loan, Matrix Computations (John Hopkins University Press, Baltimore, 2013)zbMATHGoogle Scholar
  5. 5.
    ScaLAPACK - Scalable Linear Algebra PACKage,
  6. 6.
    Matrix Algebra on GPU and Multicore Architectures,
  7. 7.
  8. 8.

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Max Planck Computing and Data FacilityGarching bei MünchenGermany

Personalised recommendations