Skip to content

GitLab

Commit 3c71b36b authored by Müller, Felix's avatar Müller, Felix
Browse files

Fix directories in Doxyfile 

Replace old src/amdis/... paths with amdis/... in Doxyfile
Build doc for all backends, not just MTL
Split howto.md into two files and include them in doc
Hide Dune namespace from doc (contains almost only forward decl)
parent 5c014a3d
Hide whitespace changes
Inline Side-by-side
Showing with 80 additions and 71 deletions
docs/Doxyfile
View file @ 3c71b36b
......@@ -122,29 +122,32 @@ WARN_LOGFILE =
#---------------------------------------------------------------------------
FILE_PATTERNS = *.hpp *.cpp *.md
INPUT = ../src/amdis \
../src/amdis/common \
../src/amdis/functions \
../src/amdis/gridfunctions \
../src/amdis/linearalgebra \
../src/amdis/linearalgebra/mtl \
../src/amdis/localoperators \
../src/amdis/operations \
../src/amdis/typetree \
../src/amdis/utility
INPUT = ../amdis \
../amdis/common \
../amdis/functions \
../amdis/gridfunctions \
../amdis/io \
../amdis/linearalgebra \
../amdis/linearalgebra/eigen \
../amdis/linearalgebra/istl \
../amdis/linearalgebra/mtl \
../amdis/linearalgebra/petsc \
../amdis/localoperators \
../amdis/operations \
../amdis/typetree \
../amdis/utility
INPUT_ENCODING = UTF-8
RECURSIVE = NO
EXCLUDE = ../src/amdis/linearalgebra/eigen \
../src/amdis/linearalgebra/istl \
../src/amdis/linearalgebra/petsc
EXCLUDE =
EXCLUDE_SYMLINKS = NO
EXCLUDE_PATTERNS =
EXCLUDE_SYMBOLS = AMDiS::Impl \
AMDiS::Math::Impl_ \
AMDiS::Concepts::Impl_ \
AMDiS::detail \
Dune \
itl::details
EXAMPLE_PATH = examples
EXAMPLE_PATH = ../examples
EXAMPLE_PATTERNS = *
EXAMPLE_RECURSIVE = NO
IMAGE_PATH =
......@@ -168,7 +171,7 @@ SOURCE_TOOLTIPS = YES
USE_HTAGS = NO
VERBATIM_HEADERS = YES
CLANG_ASSISTED_PARSING = NO
CLANG_OPTIONS = -std=c++14
CLANG_OPTIONS = -std=c++17
#---------------------------------------------------------------------------
# Configuration options related to the alphabetical class index
......
docs/tutorials/backup-restore.md 0 → 100644
View file @ 3c71b36b
# How to backup and restore problem data?
In very long simulations, i.e. many timesteps, it is recommended to backup an
intermediate state from time to time, so that it is possible to resume a
simulation from the last stored backup file. Therefore, you mostly need
a representation of the grid in memory incorporating the grid hierarchy and
the data for the solution vector. Other parameters are fixed or problem specific,
so need to be stored manually.
In the `ProblemStat` there is a pair of functions `backup()` and `restore()`
that can be used for this purpose. The backup function saves the grid and the
solution vector to files and the restore function initializes a new problem from
stored files.
```c++
ProblemStat prob("prob");
prob.initialize(INIT_ALL);
AdaptInfo adaptInfo("adapt");
// some calculation ...
prob.backup(adaptInfo);
```
This create at least two files, one for the grid and one for the solution, where
the filenames are either fixed to `backup_TIMESTEP.grid` and `backup_TIMESTEP.solution`
or specified in the initfile as
```
prob->backup->grid: backup_xyz.grid
prob->backup->solution: backup_xyz.solution
```
To restart a simulation it is recommended to initialize the problem directly with
the grid and solution file, i.e.
```c++
ProblemStat prob("prob");
prob.restore(INIT_ALL);
// some more calculation ...
```
where the restore filenames **must** be given in the initfile as
```
prob->restore->grid: backup_xyz.grid
prob->restore->solution: backup_xyz.solution
```
docs/howto.md → docs/tutorials/block-mat-vec.md
View file @ 3c71b36b
# How To
# How to use block matrices and vectors?
[TOC]
## How to use block matrices and vectors?
When using the default `BasisCreator` provided by AMDiS, like `LagrangeBasis` or
`TaylorHoodBasis`, the resulting indexing scheme is *flat*, meaning we need to
use one (sparse) matrix and one vector to describe the linear system. Sometimes
......@@ -67,51 +64,4 @@ auto& vec_vel_y = vec_vel[1];
Note: Currently the blocking is implemented for the ISTL backend only.
## How to backup and restore problem data?
In very long simulations, i.e. many timesteps, it is recommended to backup an
intermediate state from time to time, so that it is possible to resume a
simulation from the last stored backup file. Therefore, you mostly need
a representation of the grid in memory incorporating the grid hierarchy and
the data for the solution vector. Other parameters are fixed or problem specific,
so need to be stored manually.
In the `ProblemStat` there is a pair of functions `backup()` and `restore()`
that can be used for this purpose. The backup function saves the grid and the
solution vector to files and the restore function initializes a new problem from
stored files.
```c++
ProblemStat prob("prob");
prob.initialize(INIT_ALL);
AdaptInfo adaptInfo("adapt");
// some calculation ...
prob.backup(adaptInfo);
```
This create at least two files, one for the grid and one for the solution, where
the filenames are either fixed to `backup_TIMESTEP.grid` and `backup_TIMESTEP.solution`
or specified in the initfile as
```
prob->backup->grid: backup_xyz.grid
prob->backup->solution: backup_xyz.solution
```
To restart a simulation it is recommended to initialize the problem directly with
the grid and solution file, i.e.
```c++
ProblemStat prob("prob");
prob.restore(INIT_ALL);
// some more calculation ...
```
where the restore filenames **must** be given in the initfile as
```
prob->restore->grid: backup_xyz.grid
prob->restore->solution: backup_xyz.solution
```
docs/tutorials/tutorials.md
View file @ 3c71b36b
......@@ -12,27 +12,34 @@ functions.
Create a problem class as a container for the grid, basis, solution vector and
linear system and many more. What can a problem class do for your workflow.
3. [Operators](../reference/Operators.md):
3. [Backup and Restore](backup-restore.md)
Write intermediate states of your lengthy conmputations to file and resume the
simulation from a backup file.
4. [Operators](../reference/Operators.md):
Define your PDE in terms of operator terms. Learn how to add operators to a
problem and how to define your own operators.
4. [Boundary Conditions](boundary-conditions.md):
5. [Block Matrices and Vectors](block-mat-vec.md)
Use block matrices and vectors instead of the default flat versions.
6. [Boundary Conditions]():
Select parts of the grid boundary and assign boundary conditions or boundary
operators. Learn how to specify the boundary parts and how to set different
types of boundary conditions, like Dirichlet or periodic conditions, or
Neumann/Robin-type boundary integrals.
5. [Linear Solvers](linear-solvers.md):
7. [Linear Solvers]():
Control the solution of the assembled linear system, which linear solver to use
and which preconditioner. How to control solver parameters and how to add an
own linear solver.
6. [Adaptivity and Grid Refinement](grid-adaptivity.md):
8. [Adaptivity and Grid Refinement]():
Grid adaptivity needs control of marking and refinement strategies. Learn how to
define and add a marker to the problem, learn about interpolation during grid
adaptivity and how to load balance your grid.
7. [FileWriters](filewriters.md):
9. [FileWriters]():
Output and backup/restore of data is an important topic for simulations and
for postprocessing. Different Writers are introduced and the configuration and control
of these writers discussed.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment