Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Maybe isolate transport part from the TransportBase<dim> out? #10

Closed
wzheng21 opened this issue Jul 12, 2017 · 3 comments
Closed

Maybe isolate transport part from the TransportBase<dim> out? #10

wzheng21 opened this issue Jul 12, 2017 · 3 comments
Labels
major Involves multiple classes across multiple top level namespaces. question

Comments

@wzheng21
Copy link
Collaborator

The initial structure gives hopes for future functionality adding in. But the assemble part of the transport kernel is buried in a bunch of irrelevant stuffs, e.g. deal.II objects, outputs etc.

A better idea is to name current TransportBase as sth like "Carrier". The intuition is current TransportBase actually serves as an aircraft carrier carrying a lot of different "aircrafts", instead of the supposedly base class of transport methods). After the renaming, a new TransportBase will be established as a pure base class of transport methods.

Purpose: gradually cut the sloppiness; more crystal clear for better modularization; better understandability of the carrier functionality (as it only serves as non-transport stuffs).

Expecting changes are:

  1. ModelManager will be instantiated on deck of the Carrier class and build function will be run to cast the pure TransportBase to derived transport methods, such as EvenParity. Currently, this step is in the main function.
  2. We will have to change the scope of some FE objects, such as quadrature. It is currently declared in the TransportBase (the future Carrier). I expect those objects to be declared in the pure TransportBase
  3. Change the main function to instantiate Carrier instead of calling ModelManger as we will no longer explicit build a transport model outside Carrier anymore. We rather instantiate Carrier and run it.

Other impacts in near-term to long term:
TransportBase and future NDABase will have different FE objects (quadratures etc) inside corresponding classes. This makes sense because correction in NDA needs more accurate quadrature to integrate.

To this point, this gives me the impression that it sounds like a doable but non-trivial structural surgery.

What do you think? @rachelslaybaugh
@wzheng21 wzheng21 changed the title Maybe isolate transport part of the TransportBase<dim> out? Maybe isolate transport part from the TransportBase<dim> out? Jul 12, 2017
@wzheng21 wzheng21 added the major Involves multiple classes across multiple top level namespaces. label Jul 13, 2017
@wzheng21
Copy link
Collaborator Author

Note after planning meeting: use BARTDriver as the name of the manager.

@wzheng21
Copy link
Collaborator Author

This issue will be resolved after #20 but before #5

wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 24, 2017
Added bart_builder.h SlaybaughLab#10
734385f 
1. Added bart_builder.h and .cc. bart_build contains functions we
can use to project base classes to specific models.
2. Explicitly use PETScWrappers namespace instead of redefining one
as LA
3. Changed main.cc so we can build the whole thing through BartDriver
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 24, 2017
main.cc changed SlaybaughLab#10
bdb941e 
main.cc is adapted to using BartDriver<dim>
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 25, 2017
Cleaning and sorting things up gradually SlaybaughLab#10
aed7808
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Cleaning BartDriver SlaybaughLab#10
12f8217
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Cleaning BartDriver SlaybaughLab#10
42724c0
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Cleaning BartDriver SlaybaughLab#10
f259461
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Tiding TransportBase SlaybaughLab#10
4749172
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Changed radio functions directly based on MPI SlaybaughLab#10
09f7aa0
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Changed radio functions directly based on MPI SlaybaughLab#10
2e55b30
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Change PreconditionerSolver constructor SlaybaughLab#10
81dea79
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 26, 2017
Cleaning and renaming SlaybaughLab#10
83fd5b9
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 27, 2017
BartDriver cleaning SlaybaughLab#10
13f0e12
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 27, 2017
Changed output filename to a shorter version SlaybaughLab#10
58b3078
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 27, 2017
Modified builders in bart_tools.cc and .h SlaybaughLab#10
2c24803
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 27, 2017
Changed aq and builders SlaybaughLab#10
c3b144b
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 27, 2017
In progress of cleaning iteration base SlaybaughLab#10
35da9a0
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Jul 28, 2017
Cleaning continued SlaybaughLab#10
9e3aea4
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 3, 2017
Modifying transport base and derived classes to take vector input Sla…
169586f 
…ybaughLab#10
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 3, 2017
Preparing iteration base class SlaybaughLab#10
469a265
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 3, 2017
Preparing iteration base class SlaybaughLab#10
0046d93
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 5, 2017
Change TransportBase to EquationBase SlaybaughLab#10
1bdf3d8
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 7, 2017
Cleaning and restructuring SlaybaughLab#10 doxygen learning in progress
2cfce70 
SlaybaughLab#17
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
Moved tensor_norm from LSGC to EvenParity SlaybaughLab#10
827f339
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
Cleaned up bart_driver.h SlaybaughLab#10
0d1fa83
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
Modified BartDriver so vectors don't live in it. SlaybaughLab#10
83f96c6
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
Modified BartDriver so vectors don't live in it. SlaybaughLab#10
8f2c486
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
Modified BartDriver so vectors don't live in it. SlaybaughLab#10
ebce098
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
IterationBase changed to Iterations SlaybaughLab#10
fd778c9 
This class does not serve as base class for transport iterations,
so it is renamed to Iterations which is used to perform different
iterations defined by upcoming classes.
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
misc, see below SlaybaughLab#10 SlaybaughLab#17
d4ae46d 
Moved doxygen commands to header files#17
Changed other things for SlaybaughLab#10
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 10, 2017
In process SlaybaughLab#10
87e9291
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 16, 2017
Modified IterationBase as discussed SlaybaughLab#10
f9b0afe
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 21, 2017
In progress of changing structure SlaybaughLab#10
f0e4705
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 24, 2017
A commit before Marissa starts SlaybaughLab#10
e181eca
wzheng21 pushed a commit that referenced this issue Aug 24, 2017
Merge pull request #25 from weixiong-zheng-berkeley/devel
6921284 
A commit before Marissa starts #10
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Aug 29, 2017
Another fix in EquationBase SlaybaughLab#10
096fc01
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 2, 2017
Iteration class change to be incorporated in (in progress) SlaybaughL…
c697ce7 
…ab#10
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 2, 2017
Delete redundant function arguments after data structure change Slayb…
6a655db 
…aughLab#10
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 6, 2017
Adapt the code to using new structure SlaybaughLab#10
2439eda
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 6, 2017
Prototype PI as example for future iterations SlaybaughLab#10
5a3237c
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 6, 2017
EquationBase instance contains its own local cell iterators Slaybaugh…
937116a 
…Lab#10

This is a major design change. It means we will initialize
cell iterators on current process when we instantiate an equation.
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 9, 2017
All the changes made to make SlaybaughLab#10 work (in progress)
1f70d33
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 13, 2017
Renamed function parameters to keep readability SlaybaughLab#10
b4d42ee
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 13, 2017
Coordinate iteration related classes SlaybaughLab#10
d2d164f
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Sep 13, 2017
Gauss-Sidel iteration SlaybaughLab#10
2cfc091
@wzheng21 wzheng21 mentioned this issue Sep 14, 2017
Merged
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Oct 26, 2017
Debugging in progress SlaybaughLab#10
8a6e11d 
In infinite medium one-group eigen problem, it gives correct flat
results;

Fixed the issue of non-stopping execution that code goes to 3d after
2d calculations.

Made effort to reduce warning in building process
wzheng21 pushed a commit to wzheng21/BART that referenced this issue Oct 27, 2017
Debugging in progress SlaybaughLab#10
ea96c47 
In infinite medium one-group eigen problem, it gives correct flat
results;

Fixed the issue of non-stopping execution that code goes to 3d after
2d calculations.

Made effort to reduce warning in building process
@wzheng21 wzheng21 mentioned this issue Oct 27, 2017
Merged
@wzheng21
Copy link
Collaborator Author

Up to this point, structure change is complete.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
major Involves multiple classes across multiple top level namespaces. question
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@wzheng21