11.08.15 netcdf cont - daplff/hiWi_cont GitHub Wiki

todo:

  • test running C-bound netcdf4 test program
  • implement netcdf output class in program

log

  • Started testing with simplxy examples from netcdf tutorial, works well, need linking flag -lnetcdf
  • Tested class, works as well, provided no nans/infs in output
  • netcdf output works! however, it currently produces data files that are bigger than the raw text files. Not really sure what to think about this, but I guess I'll experiment with that later. NOTE IT THOUGH

what's the next step again?

checking back on yesterday's log: (copypaste, ping)

  • netcdf (big part done!)
  • result validations
  • direct I/O to fortran without extra copy (possible?)
  • startup parameters from class (-idea! have basic/default tree, load it or a custom one, then upon wish edit entries)
  • ofc., extend program for more simulator variables in C++.

so...

  • try to go on to visualising netcdf files - included ypos as well

  • paraview reads netcdf, but apparently not in the format I wrote - could be investigated

  • but: tried for too long trying to install python-netcdf, not really any use trying further unless new approach is tried

  • BUT: now got it to work, by not using import NetCDF or something, but from Scientific.IO.NetCDF import *

for tomorrow:

either

  • look into other way of loading into paraview ** don't need python for going from netcdf4->vtk
  • go for other issue: somthing like
  • cleaning up file structure & docu
  • validation (important)
  • input parameters and stuff

discussables with Roland:

  • netcdf bigger than ascii?
  • installation of python-netcdf4 (maybe obsolete now that i found it in Scientific.IO - that one doesn't have netcdf4 though)