Open Hub
Open Hub

 Settings |  Report Duplicate

1
I Use This!
Very Low Activity

Commits : Listings

Analyzed 14 days ago. based on code collected 14 days ago.
Jan 29, 2024 — Jan 29, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Eliminated NCTool::var_to_da_vec() and associated VecScatter to processor zero and integer indexing scheme. All functionality replaced by that in regrid_local|global_var(), which now works for arbitrary spacing in the vertical. Note regrid_global_var() now has ability to regrid an integer mask using a MaskInterp tool. Also: attempted to assure ANSI C compatibility for all .c source files; attempted, by adding CXXLINKER to Makefile, to help fix build errors associated to PETSc linker-choice bug and derived PISM bug #9371.
bueler
as Ed Bueler
More... about 17 years ago
Regridding implemented for non-equally spaced vertical grids. __PLEASE TRY TO BREAK IT AND TELL ME ABOUT IT___ It remains to remove var_to_da_vec() from bootstrapping, replacing with regrid_local_var().
bueler
as Ed Bueler
More... about 17 years ago
Comments regarding parallel regridding and irregular grid.
Jed Brown
More... about 17 years ago
Bug fix revealed by test E: communication of vWork2d[] arrays had to happen sooner, within velocity(), for basalSIA().
bueler
as Ed Bueler
More... about 17 years ago
I think unequal spaced grid works well (i.e. no better than lots of equal spaced levels, but not much worse, and with some speedup), except: regridding and Test K results. IceParam *p removed from IceGrid; removes 'p->' from everywhere; remember PetscBag? iMlegacy.cc no longer even compiled.
bueler
as Ed Bueler
More... about 17 years ago
Unequally-spaced vertical implementation still not stable. Verification in Test G pretty good. Init -if and regrid still need work. Now removing legacy code as too hard to maintain.
bueler
as Ed Bueler
More... about 17 years ago
Builds. Test K and Test F results not great but not silly.
bueler
as Ed Bueler
More... about 17 years ago
Option -cheb now implemented; produces vertical grid spaced like the [-1,0] portion of the Chebyshev extreme points. Aspects of temperature solution, with and without bedrock, still need addressing. Test K verification not so good yet. Note bootstrap stuff separated into src/iMbootstrap.cc. Note src/iMIOnetcdf.cc moved into src/iMIO.cc. Regridding for unequal spaced grids not yet implemented. THIS REVISION DOES NOT LINK right now because IceModelVec3Bedrock::getValColumn() and setValColumn() are not yet implemented.
bueler
as Ed Bueler
More... about 17 years ago
Ongoing structural cleanup and documentation related to non-equally-spaced vertical grid and Vec abstraction. Of the numerical models, only the temperature temperatureStep() uses an internally equally spaced grid at this point; it then linear interpolates to the storage grid. This can be changed later if desired; verification and profiling should determine what happens. Verification shows that the linear interpolation is doing no harm if the storage grid actually has equal spacing.
bueler
as Ed Bueler
More... about 17 years ago
Bit more work on removing equal spacing dependence. Revised and documented grain size model.
bueler
as Ed Bueler
More... about 17 years ago
Abstraction of 3d Vecs now in good shape. Still equally spaced. Side effects have been: (1) some slow-down from lots of linear interpolation code and (2) some multiplexing of communication because each 3d Vec has its own DA and there is no conflict if we try to begin several and then end several LocalToLocals.
bueler
as Ed Bueler
More... about 17 years ago
Critical fix to make -Mz option work. Note Test G is not performing as well as in pre 235 revisions.
bueler
as Ed Bueler
More... about 17 years ago
Abstraction of 3D Vecs mostly completed, including former vWork3d.
bueler
as Ed Bueler
More... about 17 years ago
First implementation of abstraction of 3d Vecs now works. Some slowdown with current fine equally-spaced vertical grid, of course; no actual implementation of non-equal grid yet. verifynow.py essentially same output.
bueler
as Ed Bueler
More... about 17 years ago
Major revision mostly done; won't build yet. 3d Vecs are abstracted, in preparation for changing to unequally-spaced grid. Still need to complete in these ways: (1) iMvelocity.cc, (2) all uses of vWork3d[], (3) NetCDF read and write, (4) communication simultaneous with copying.
bueler
as Ed Bueler
More... about 17 years ago
Added horizontal ice flux to PISM output NetCDF files. A bit more work on MISMIP. A start on abstracting da3-based Vecs.
bueler
as Ed Bueler
More... about 17 years ago
Initial implementation of MISMIP: not to be trusted until more seriously tried. Needs explicit treatment of calving front. Needs text output routines. Needs management scripts.
bueler
as Ed Bueler
More... about 17 years ago
PDD model probably done; see trials in comment at end of src/base/iMpdd.cc. *Monte carlo method produces results in close agreement with (Calov and Greve 2005). Note Calov-Greve clearly needs more than monthly evaluations for quadrature; at this point we use weekly. PDD re-documented in manual. Maintenance. Superficial start on MISMIP; not functional at all.
bueler
as Ed Bueler
More... about 17 years ago
Debugging associated to Calov-Greve PDD model. Maintenance.
bueler
as Ed Bueler
More... about 17 years ago
Modifications to summary line: step moved to flag line. Initial implementation of (Calov and Greve 2005) positive degree day scheme; not trustable without additional testing.
bueler
as Ed Bueler
More... about 17 years ago
Ongoing fiddles to plastic till mod of EISMINT II exper I. Work on manual.
bueler
as Ed Bueler
More... about 17 years ago
Fixed and simplified and improved docs on regrid.
bueler
as Ed Bueler
More... about 17 years ago
Fiddlesticks.
bueler
as Ed Bueler
More... about 17 years ago
Update to plastic till experiment script: add bedrock thermal.
bueler
as Ed Bueler
More... about 17 years ago
Update to plastic till experiment script.
bueler
as Ed Bueler
More... about 17 years ago
I believe I have completed the conversion to CISM names for NetCDF variables. Units still do not match (in general). Modifications of the plastic till SSA EISMINT II experiment I code.
bueler
as Ed Bueler
More... about 17 years ago
Switched to GLIMMER names for variables in PISM NetCDF files. Units not changed because we conform closer to CF-1.0 than GLIMMER, I think. Also changes to tools and manual. Need to modify scripts for EISMINT-Greenland and EISMINT-Ross to use new variable names.
bueler
as Ed Bueler
More... about 17 years ago
Small fix to test/eis2Iplastic.sh.
bueler
as Ed Bueler
More... about 17 years ago
Improved reporting and time series extraction. Manual needs revision accordingly ...
bueler
as Ed Bueler
More... about 17 years ago
Script modifications: (i) fix of fill_missing.py to fix for old pycdf(?); (ii) clarification of eis2Iplastic.sh.
bueler
as Ed Bueler
More... about 17 years ago