| [pism] / trunk / src / verif / exactTestsFG.h |
Cleanup of documentation, especially comments and references for exact solutions.
Finalized removal of Ruby dependecy (ncgen.rb). Converted Sigma to J/(s m^3) units; before was K/s. New units compatible with simple formula Sigma = D_ij tau_ij, and with new paper. (Old units appeared in Bueler et al 2007 J. Glaciol.)
Cleaned up viewer sizing and standard out summary. Simplified src tree structure slightly; it is actually useful to have some structure, but I had overdone it.
Fairly widespread changes:
0. Makefile changed. Now "make" behaves correctly w.r.t. modified files, and "make install" puts
libraries in lib/ and executables in bin/ and cleans up .o and .d files. Manual reflects this.
1. Change to pism_state.cdl: Made climate variables also depend on t (task #5200). Made t(t) a double
(because we see effects of single precision all the time in long runs). Added pism_intent
attribute where known; values are model_state, climate_steady, climate_snapshot, diagnostic.
See comment near start of pism_state.cdl.
2. Modified nc_util etc to correctly pass around a double t, including for regrid.
3. iMregrid.cc now contains netcdf version (the default) and the unsupported (and probably not functional)
old .pb version. iMregrid_netcdf.cc deleted. SHOULD WE GO THROUGH AND DELETE RELIC .pb STUFF?
4. bootstrapFromFile_netcdf() has different reporting on defaults. It should be fairly robust w.r.t.
different kinds of -bif files.
5. small possible temperature bug changed: use ice material properties in k=0 segment of column if
Mbz==1
6. setSoundingFromOptions() put in afterInitHook(); drivers changed to remove explicit call
7. improved the viewer code; note addition of utility functions getSurfaceValueOf3D() and
getSliceOf3D(); note ideas in task #4989 still probably worth adding
8. Added better error reporting of pismv -test F|G: surface velocities and Sigma; does task #4184.
THIS REVEALS test G DOES NOT CONVERGE FOR SURFACE VELS OR FOR Sigma, AS SUSPECTED FROM EISMINT II
TESTS. This better reporting should help with implementing margin tricks (task #5244).
9. Small edits to iceGRNModel.cc. Changed set from options logic re enhancement factor.
Added source sub-directory for exact tests.
reportErrors() in IceCompModel improved, including for test E; storage and viewing of basal velocities partly implemented; other small changes
moved exact tests to C (from C++) for compatibility; updates to Makefile
Powered by ViewCVS