1. 17 May, 2013 1 commit
  2. 16 May, 2013 1 commit
    • Pierre Cazenave's avatar
      Make the interpolation of the regularly gridded data to the unstructured grid... · 2980dfc0
      Pierre Cazenave authored
      Make the interpolation of the regularly gridded data to the unstructured grid occur in parallel. So far as I can tell (and have tested) this has not changed the results of the interpolation, but the speedup is significant (from almost interminable to over in the blink of an eye). Getting the loop to be able to be parallelised required refactoring the code a bit (mainly through the creation of intermediate arrays), but as I said, it seems to have improved things speed-wise
      2980dfc0
  3. 14 May, 2013 1 commit
  4. 14 Feb, 2013 1 commit
  5. 08 Feb, 2013 1 commit
  6. 04 Feb, 2013 2 commits
  7. 09 Jan, 2013 1 commit
  8. 14 Nov, 2012 1 commit
  9. 13 Nov, 2012 1 commit
  10. 09 Nov, 2012 1 commit
  11. 08 Nov, 2012 1 commit
  12. 07 Nov, 2012 1 commit
  13. 02 Nov, 2012 1 commit
    • Pierre Cazenave's avatar
      Move the code to write the NetCDF file into a new function... · ca7e2f1f
      Pierre Cazenave authored
      Move the code to write the NetCDF file into a new function (write_FVCOM_forcing.m). Also improve the speed of the interpolation in two ways: 1. Use TriScatteredInterp to do the interpolation (with natural neighbour interpolation), and 2. Do the interpolation onto the element centres using their positions directly, rather than averaging the nodal values for each element. This last step significantly improved the speed of the interpolation
      ca7e2f1f