1. 08 May, 2018 2 commits
  2. 07 May, 2018 1 commit
  3. 04 May, 2018 1 commit
  4. 03 May, 2018 2 commits
    • Pavel Kus's avatar
      copy a_dev to a_mat at the end of bandred instead of later · 60b33fe9
      Pavel Kus authored
      Previously, matrix a has been kept on the device at the end of bandred
      and then copied to host later in redist_band (called from tridiag_band).
      However, it seems to be more convenient to do it at the end of bandred
      (since it has to be done anyway). Band to tridi is thus now not using a
      on the device at all.  However, we keep it in the interface, sice it
      might be usefull in the future.
      60b33fe9
    • Pavel Kus's avatar
      renaming matrix a to a_mat · 830b54aa
      Pavel Kus authored
      830b54aa
  5. 26 Apr, 2018 1 commit
    • Pavel Kus's avatar
      tile size can be influence by the user · 0183d8ca
      Pavel Kus authored
      tile size is used in elpa1_tridiag and elpa2_bandred.
      Iits value is computed based on block size and least common multiple of
      number of processes rows and colums, but then can be enlarged to some
      multiple of this value. How much it should be enlarged can now be
      influenced by min_tile_size parameter.
      0183d8ca
  6. 12 Jan, 2018 1 commit
  7. 11 Jan, 2018 2 commits
  8. 14 Dec, 2017 1 commit
  9. 02 Oct, 2017 2 commits
  10. 13 Sep, 2017 1 commit
  11. 11 Sep, 2017 1 commit
  12. 29 Aug, 2017 1 commit
  13. 26 Aug, 2017 2 commits
  14. 25 Aug, 2017 3 commits
  15. 24 Aug, 2017 1 commit
  16. 03 Aug, 2017 1 commit
  17. 30 Jul, 2017 1 commit
  18. 28 Jun, 2017 2 commits
  19. 26 May, 2017 1 commit
  20. 22 May, 2017 1 commit
  21. 21 May, 2017 1 commit
  22. 27 Apr, 2017 1 commit
  23. 06 Apr, 2017 1 commit
  24. 23 Mar, 2017 1 commit
  25. 14 Feb, 2017 2 commits
  26. 13 Feb, 2017 1 commit
    • Andreas Marek's avatar
      Error in "error handling" · 78543e3c
      Andreas Marek authored
      It could happen that ELPA stopped with an error but an exit code "0"
      was given, i.e. one could assume everything was fine when it was not!
      Now each Fortran "stop" was replaced with "stop 1" to prevent this
      78543e3c
  27. 10 Feb, 2017 1 commit
  28. 07 Feb, 2017 1 commit
  29. 06 Feb, 2017 2 commits
  30. 05 Feb, 2017 1 commit