Wiki
Clone wikipflotran / Developers / CodeDevelopment / DevTeleconNotes
Telecon - next
- Discuss how reactive_transport_auxvar_type%auxiliary_data can be used to store data (e.g. in reaction sand boxes, etc.
Telecon - 05/11/2016
- Need to resolve documentation issue
- Consider github+travis-ci
Telecon - 02/24/2016
- PFLOTRAN workshop/developer meeting
- Logistics
- Poll user base to determine interest
- STOL (e.g. for Richards in domains with high pressures)
- FORD: https://github.com/cmacmackin/ford
Telecon - 02/10/2016
- File and object naming convention
- Refactoring waypoints * Move FINAL_TIME to simulation block * Move CHECKPOINT to simulation block
- Wrapping HDF5 I/O. Developing a PFLOTRAN middle layer to avoid all the preprocessing whenever SCORPIO or raw HDF5 is used.
Telecon - 03/25/2015
- File and object naming convention
Telecon - 03/11/2015
- Discuss [CLM-PFLOTRAN coupling](https://bitbucket.org/pflotran/pflotran-dev/wiki/Documentation/CodeDevelopment/CLM-PFLOTRAN-Notes)
Telecon - 01/28/2015
- Attendees:
- Input/Init refactor
- Discuss [MINC](https://bitbucket.org/pflotran/pflotran-dev/wiki/Documentation/CodeDevelopment/MINC2) in RICHARDS (Gautam)
- TS and adjoint:
Jed Brown writes: TS works with the formulation
f(u,udot) = g(u)
where f may be stiff and g must be non-stiff. When using a method that involves linear solves, you provide an (approximate) Jacobian
J_a(u,udot) = ∂f/∂u + a ∂f/∂udot
where the parameter a is defined by PETSc (depends on the integration method and step size).
PCL: This can be important in multicomponent systems where A(u1, u2, …, uN) is a function of N variables. In the latter case the Jacobian involves the first derivative of A whereas in the former it involves the second derivative and is much more complicated.
JB: Indeed, I wanted to avoid that.
Telecon - 01/14/2015
- Attendees: Glenn, Gautam, Peter, Satish, Richard
- Discuss MINC in RICHARDS (Gautam)
- Init()/input file refactor (Glenn)
- Multicomponent gas diffusion (Peter)
- Interest for block grid refinement capability in PFLOTRAN (by Paolo)
Telecon - 10/22/2014
- Discuss output of input file contents
- Discuss removal of MFD, Immiscible/miscible
Telecon - 10/01/2014
- Discuss dataset formats and how to streamline
Telecon - 04/16/2014
- Reducing 3D problem to a 1D (Gautam)
Telecon - 03/05/2014
- Outstanding pull request for using material_aux in TH mode (Gautam)
- pull request 69 - Fengming's changes for mira - bja: I looked at the makefile changes and I think there are bugs. I can't use Fengming's petsc install so I'm building my own. I think the best path forward (least work for pflotran-dev) is to reapply the patch on a clean clone, fix the errors and commit giving Fengming credit.
- init refactor - bja: needs to be done in stages....
- provenance - bja
- unit tests - bja
- thermodynamic database format and support for different databases - pcl
Telecon - 01/08/2014
- TH(C) duplication - remove THC or consolidate TH/THC into a single implementation?
- PM refactor Phase 2 (permanently disable legacy version)
- refactor of init
- refactor of regression
- consolidating *_Simulation_class (single simulation class with pmc trees providing the differences)
- transfer Residual/Jacobian/Other subroutines from MODE.F90 to pm_MODE.F90. [This would possibly require Ctrl-C/V] (bja: Does this get rid of the factory and HijackSimulation stuff, or is that a separate TODO? gb: HijackSimulation/HijackTimestepper will no longer be needed with the init refactor. *_factory.F90 may also be consolidated with the conslidation of *_Simulation_class.)
- move global_aux.F90 into simulation_aux.F90
- Unit testing: Are we ready to start?
- Documentation
- function headers
- evaluate OpenGrok: pflotran.lbl.gov
Meeting at AGU - 12/2013
- [HDF5 dataset refactor](https://bitbucket.org/pflotran/pflotran-dev/wiki/Documentation/CodeDevelopment/HDF5DatasetRefactor) (Gautam)
- Consolidation of PatchGetVariable1() and PatchGetVariableValueAtCell() (Gautam)
Telecon - 11/27/2013
Telecon - 11/13/2013
- Regression testing for standard output files (.h5, .tec, observation, mass balance, etc.)
- [HDF5 dataset refactor](https://bitbucket.org/pflotran/pflotran-dev/wiki/Documentation/CodeDevelopment/HDF5DatasetRefactor) (Gautam)
- Modification of BC/SS for TH+ice MODE (Gautam: Create a pflotran.in to reproduce the issue)
- Provenance information compiled into pflotran executable, always dumped to screen, output files? (Ben)
- Regression testing for surface flows (input refactor dependent) (Ben)
- FAQ page including [undefined PETSC_* env variables](https://groups.google.com/forum/#!topic/pflotran-users/mS6oWiTKvgA) and how to productively request help / report a bug (include hg parent, hg status info instead of stating "latest pflotran", petsc version info). (Ben)
Telecon - 10/30/2013 [Was cancelled]
** Regression testing for standard output files (.h5, .tec, observation, mass balance, etc.)**
Telecon - 10/16/2013
- [Example of mixed case input.](https://bitbucket.org/pflotran/pflotran-dev/wiki/Documentation/CodeDevelopment/InputFileRefactor) (e.g. Final_Time instead of FINAL_TIME) (Ben)
- Example input file with END CARD (Glenn)
Telecon - 10/02/2013
- Switch makefile -> makefile_legacy and makefile_new -> makefile (Glenn - completed 10/2/13)
- Date for deprecating old code is Nov 30, 2013 (Glenn)
- Rename regression test directories (Glenn - completed 10/2/13)
- Refactoring init.F90 in experimental branch (Ben/Glenn)
- Verify accuracy mphase process model approach (Peter/Satish)
Telecon - 9/04/2013
- Use waypoint for checkpointing.
Telecon - 8/21/2013
- Add pointer to simulation_aux object from PMCs instead of passing as argument (requested by many, to be implemented by Gautam)
- Always define CHUAN_CO2 in the makefile so that CO2 regression tests can be run (requested by Peter, to be implemented by Glenn - completed 10/2/13)
- Rename Synchonize1/2/3() routines and PMC (requested by many, to be implemented by Ben/Gautam/Glenn)
- Add ability to accumulate properties (e.g. mass) between time steps. Similar to SynchronizeXXX, but accumulates instead of updates (requested by Gautam, to be implemented by ?)
- Restructuring of input file to handle input for multiple PMs with differing grids, etc. Need to develop prototypes of input file formats and submit to group for review.
- Refactor regression testing portion of code to not be specific to subsurface. This will require refactoring all the code in Init(). (requested by Ben, to be implemented by Ben/Glenn)
- Revise PFLOTRAN website (requested by Satish, to be updated by all)
- Add F2003 pointer to function capability to redirect reactive transport flux calculations to code specific to uniform flow and transverse dispersion ( requested by Peter, to be implemented by ?)
Updated