NEWS    LOBPCG method of Andrew Knyazev used on the K computer in Japan for superconductivity research

Date:

March 20, 2018

Where:

Asian Conference on Supercomputing Frontiers

Description:

RIKEN is Japan's largest comprehensive research institution. The RIKEN Center for Computational Science (R-CCS) is the leadership research center in high performance computing and computational science in Japan, operating the K computer in Kobe, Japan, which achieves 602.7 teraflops on the High Performance Gradients (HPCG), making it the world HPCG top.

R-CCS researchers and collaborators in their new paper use the Hubbard model for strongly-correlated electron systems to understand the mechanism behind the superconductivity. To calculate the ground state of the Hamiltonian, they implement on the K computer and successfully test the Locally Optimal Block Preconditioned Conjugate Gradient (LOBPCG) method, invented in 2001 by Andrew Knyazev; see his recent review of LOBPCG.

External Link:

https://link.springer.com/content/pdf/10.1007%2F978-3-319-69953-0_14.pdf

• Related Publications

   Knyazev, A., "Recent implementations, applications, and extensions of the Locally Optimal Block Preconditioned Conjugate Gradient method (LOBPCG)", Householder Symposium on Numerical Linear Algebra, June 2017.

  BibTeX TR2017-078 PDF

  • @inproceedings{Knyazev2017jun,
  • author = {Knyazev, Andrew},
  • title = {{Recent implementations, applications, and extensions of the Locally Optimal Block Preconditioned Conjugate Gradient method (LOBPCG)}},
  • booktitle = {Householder Symposium on Numerical Linear Algebra},
  • year = 2017,
  • month = jun,
  • url = {https://www.merl.com/publications/TR2017-078}
  • }