1. 11 Feb, 2016 1 commit
    • Andreas Marek's avatar
      Enable single-precision calculations for ELPA1 · de6a4fde
      Andreas Marek authored
      With the configure option "--enable-single-precision" ELPA1 is build
      with single-precision (half-words) only.
      
      The best precision in single-precision (float or complex) is
      2^-23 ~ 1.2e-7. The accuracy of the error residual of ELPA1 in
      single-precision mode is of the order 1e-4 to 1e-5. The orthogonality of
      the EV's is fullfilled up to about ~1e-6.
      
      Thus the precision of ELPA1 in single-precision mode is roughly 100 -
      1000 times less than the best achievable precison. This is consistent
      with the double-precision mode, where also a factor of 100 - 1000 less
      precision than the theoretical best one is found.
      
      The float EVs are identical to the double EVs to at least 1e-2, the
      precision of the EVs is thus about 1e-7/1e-2 = 1e5 times lower than the
      best theoretical precision. If the same holds for the double precision
      calculations, this implies that the double precision results can also
      be only trusted on the level 1e-11 (5 orders of magnitude larger
      than the best theoretical precision)
      
      The best speed-up compared to the double precision calculation is
      a factor of two. This is by far not achieved yet, since the singl
      precision version is not at all optimized at the moment
      de6a4fde
  2. 04 Jan, 2016 2 commits
    • Andreas Marek's avatar
      Include forgotten header files · 0aa7e6be
      Andreas Marek authored
      In some C test programs the function declaration header files
      have been missing
      0aa7e6be
    • Andreas Marek's avatar
      Started to remove depecrated Fortran variable declerations · 0a05f7d3
      Andreas Marek authored
      The Fortran variable declerations "variable type*[4,8,16]" is non
      Fortran standard. It might cause problem in the future.
      Furthermore, the usage of Fortran and C togehther is more clean
      if variables are defined according to C variable types.
      
      This is done, now for all the test programs
      0a05f7d3
  3. 22 Dec, 2015 1 commit
  4. 16 Dec, 2015 2 commits
    • Lorenz Huedepohl's avatar
      Wrong argument order · 9edc719b
      Lorenz Huedepohl authored
      9edc719b
    • Andreas Marek's avatar
      Add interface to unify C and Fortran names · bb046d1c
      Andreas Marek authored
      This commit does not change the interfaces defined in ELPA_2015.11.001 !
      All functionality is available via the interface names and definitions
      as in ELPA_2015.11.001
      
      But some new interfaces have been added, in order to unfiy the
      references from C and Fortran codes:
      
      - The procedures to create the ELPA (row/column) communicators are now
        available from C _and_ Fortran with the name "get_elpa_communicators".
        The old Fortran name "get_elpa_row_col_comms" and the old C name
        "elpa_get_communicators" are from now on deprecated but still available
      
      - The 1-stage solver routines are available from C _and_ Fortran via
        the names "solve_evp_real_1stage" and "solve_evp_complex_1stage".
        The old Fortran names "solve_evp_real" and "solve_evp_complex" are
        from now on deprecated but still functional.
      
      All documentation (man pages, doxygen, and example test programs) have
      been changed accordingly.
      
      This commit implies a change in the API versioning number, but no
      changes to codes calling ELPA (if they have been already updated to the
      API of ELPA_2015.11.001)
      bb046d1c
  5. 11 Dec, 2015 2 commits
  6. 13 Nov, 2015 1 commit
  7. 03 Nov, 2015 2 commits
    • Andreas Marek's avatar
      Remove of Makefile.example · d86c8c96
      Andreas Marek authored
      A build without autotools is not officially supported anymore.
      Thus this --- broken since long time --- Makefile.example is removed
      d86c8c96
    • Andreas Marek's avatar
      Update of c test cases · 505004e7
      Andreas Marek authored
      The examples, how to invoke ELPA from a c program have been updated.
      There are now examples for ELPA1 and ELPA2 both real and complex case.
      The test cases are still with less functionality than their Fortran
      counter parts, they are just ment as a "proof-of-concept".
      505004e7
  8. 28 Oct, 2015 1 commit
    • Alexander Heinecke's avatar
      This commit improves ELPA's performance on Intel(R) Xeon(R) E5v2 and E5v3 series CPUs by: · fe63372d
      Alexander Heinecke authored
      - enabling fusing iterations of stage 5 in ELPA2 for every configuration
      - Changed reduction bandwidth in ELPA2 to be at least 64
      - partial OpenMP parallelization of the QR factorization in bandred_real
      - OpenMP parallelization of SYMM
      - OpenMP parallelization of SYR2K in bandred_real
      - OpenMP parallelization for elpa1_reduce_add_vectors and elpa1_transpose_vectors
      - AVX2 support in backtransformation elpa2_kernels (FMA3 instructions introduced with Haswell microarchitecture)
      fe63372d
  9. 16 Jun, 2015 2 commits
  10. 01 Jun, 2015 1 commit
  11. 28 May, 2015 2 commits
  12. 24 Mar, 2015 1 commit
  13. 16 Mar, 2015 3 commits
  14. 11 Mar, 2015 1 commit
  15. 09 Mar, 2015 2 commits
    • Andreas Marek's avatar
      Correct usage of MPI abort · ff61bd20
      Andreas Marek authored
      In the test programms "MPI_ABORT" has been used incorrectly.
      ff61bd20
    • Andreas Marek's avatar
      Change of get_elpa_row_col_comms to a function · fc80ad07
      Andreas Marek authored
      The procedure "get_elpa_row_col_comms" is now a function instead
      a subroutine and returns an error code.
      
      This should have been done previously when all the other ELPA
      function have been updated to return an error code.
      
      This requires an ABI change in the next ELPA release !
      fc80ad07
  16. 11 Feb, 2015 1 commit
    • Andreas Marek's avatar
      Abort on error in QR-decomposition · ff13970b
      Andreas Marek authored
      If the QR-decomposition is used wrongly (matrix size is not a
      multiple of block size) the the execution will abort, in
      order to prevent the wrong results, discussed in a previous commit
      
      Debug messages are now available by setting the environment variable
      "ELPA_DEBUG_MESSAGES" to "yes".
      ff13970b
  17. 03 Feb, 2015 2 commits
  18. 29 Jan, 2015 1 commit
  19. 27 Jan, 2015 1 commit
  20. 30 Nov, 2011 1 commit
    • Volker Blum's avatar
      VB, TA, RJ, AM: · d7a57612
      Volker Blum authored
      We have taken the unusual (one-time) step to modify the structure
      of the ELPA git repository as follows:
      
      1) The current ELPA version has been designated as a "stable"
         version. It has proven to be rock solid for us so far, and
         deserves this name.
      
      2) We have decided on a stable version naming scheme,
         "ELPA_year.month". The first such version is
         ELPA_2011.12. This version will only ever be updated
         for clear and unambiguous bug fixes.
      
      3) All further development will happen in the directory named
         ELPA_development_version. At this time, ELPA_2011.12 is an
         exact copy of ELPA_development_version.
      
      No substantial changes have been made to either repository as yet.
      (But we are happy to say that the present "ELPA" publication, which
      we hope ELPA users will cite, now has its final volume and page
      numbers.)
      
      We apologize if this change causes any inconvenience for you. Please
      try to make sure that any changes you might make really affect the
      development version, and not the stable version. We do not plan any
      structural changes of a similar scope at any point in the future,
      but it had to be done once.
      
      Happy ELPA stable version day!
      d7a57612
  21. 22 Sep, 2011 1 commit
  22. 14 Jul, 2011 1 commit
    • Werner Jürgens's avatar
      RJ: Added two programs to read (and process) matrices from files. · ab963a3e
      Werner Jürgens authored
      There are two new programs which read matrices from a file, solve the
      real (generalized) eigenvalue problem, print the eigenvalues and check
      the result (all using elpa1 only).
      
      The author of these programs (plus documentation and Makefile) is
      Rainer Johanni (RJ). The files are checked in by Werner Jürgens.
      ab963a3e
  23. 12 May, 2011 2 commits