Skip to content

Commit

Permalink
rework the unit test docs (#1570)
Browse files Browse the repository at this point in the history
We split the current page into 3 separate pages and add descriptions for all of the unit tests
  • Loading branch information
zingale authored Jul 15, 2024
1 parent 24fc854 commit 8fd0d4a
Show file tree
Hide file tree
Showing 4 changed files with 332 additions and 181 deletions.
119 changes: 119 additions & 0 deletions sphinx_docs/source/comprehensive_tests.rst
Original file line number Diff line number Diff line change
@@ -0,0 +1,119 @@
************************
Comprehensive Unit Tests
************************

Generally, for each test, you simply type ``make`` in the test
directory. There are a number of runtime parameters that can
control the behavior. These are specified (along with defaults)
in ``_parameters`` files in each test directory and can be
overridden in an inputs file or on the commandline.

Some additional details on a few of the comprehensive unit tests
are given below.

EOS test (``test_eos``)
=======================

``Microphysics/unit_test/test_eos/`` is a unit test for the equations
of state in Microphysics. It sets up a cube of data, with
:math:`\rho`, :math:`T`, and :math:`X_k` varying along the three
dimensions, and then calls the EOS in each zone. Calls are done to
exercise all modes of calling the EOS, in order:

- ``eos_input_rt``: We call the EOS using :math:`\rho, T`. this is the
reference call, and we save the :math:`h`, :math:`e`, :math:`p`, and
:math:`s` from here to use in subsequent calls.

- ``eos_input_rh``: We call the EOS using :math:`\rho, h`, to recover
the original :math:`T`. To give the root finder some work to do, we
perturb the initial temperature.

We store the relative error in :math:`T` in the output file.

- ``eos_input_tp``: We call the EOS using :math:`T, p`, to recover the
original :math:`\rho`. To give the root finder some work to do, we
perturb the initial density.

We store the relative error in :math:`\rho` in the output file.

- ``eos_input_rp``: We call the EOS using :math:`\rho, p`, to recover
the original :math:`T`. To give the root finder some work to do, we
perturb the initial temperature.

We store the relative error in :math:`T` in the output file.

- ``eos_input_re``: We call the EOS using :math:`\rho, e`, to recover
the original :math:`T`. To give the root finder some work to do, we
perturb the initial temperature.

We store the relative error in :math:`T` in the output file.

- ``eos_input_ps``: We call the EOS using :math:`p, s`, to recover the
original :math:`\rho, T`. To give the root finder some work to do,
we perturb the initial density and temperature.

Note: entropy is not well-defined for some EOSs, so we only attempt
the root find if :math:`s > 0`.

We store the relative error in :math:`\rho, T` in the output file.

- ``eos_input_ph``: We call the EOS using :math:`p, h`, to recover the
original :math:`\rho, T`. To give the root finder some work to do,
we perturb the initial density and temperature.

We store the relative error in :math:`\rho, T` in the output file.

- ``eos_input_th``: We call the EOS using :math:`T, h`, to recover the
original :math:`\rho`. To give the root finder some work to do, we
perturb the initial density.

Note: for some EOSs, :math:`h = h(\rho)` (e.g., an ideal gas), so there
is no temperature dependence, and we do not do this test.

We store the relative error in :math:`\rho` in the output file.

This unit test is marked up with OpenMP directives and therefore also
tests whether the EOS is threadsafe.

To compile for a specific EOS, e.g., helmholtz, do::

make EOS_DIR=helmholtz -j 4

Examining the output (an AMReX plotfile) will show you how big the
errors are. You can use the ``amrex/Tools/Plotfile/`` tool
``fextrema`` to display the maximum error for each variable.


Network test (``test_react``)
=============================

``Microphysics/unit_test/test_react/`` is a unit test for the nuclear
reaction networks in Microphysics. It sets up a cube of data, with
:math:`\rho`, :math:`T`, and :math:`X_k` varying along the three
dimensions (as a :math:`16^3` domain), and then calls the EOS in each
zone. This test does the entire ODE integration of the network for
each zone.

The state in each zone of our data cube is determined by the runtime
parameters ``dens_min``, ``dens_max``, ``temp_min``, and ``temp_max``
for :math:`(\rho, T)`. Because each network carries different
compositions, we specify the composition through runtime parameters in
the ``&extern`` namelist: ``primary_species_1``,
``primary_species_2``, ``primary_species_3``. These primary species
will vary from X = 0.2 to X = 0.7 to 0.9 (depending on the number).
Only one primary species varies at a time. The non-primary species
will be set equally to share whatever fraction of 1 is not accounted
for by the primary species mass fractions.

This test calls the network on each zone, running for a time
``tmax``. The full state, including new mass fractions and energy
release is output to a AMReX plotfile.

You can compile for a specific integrator (e.g., ``VODE``) or
network (e.g., ``aprox13``) as::

make NETWORK_DIR=aprox13 INTEGRATOR_DIR=VODE -j 4

The loop over the burner is marked up for OpenMP and CUDA and
therefore this test can be used to assess threadsafety of the burners
as well as to optimize the GPU performance of the burners.
11 changes: 9 additions & 2 deletions sphinx_docs/source/index.rst
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,6 @@ for astrophysical simulation codes.
data_structures
autodiff
rp_intro
unit_tests

.. toctree::
:maxdepth: 1
Expand Down Expand Up @@ -53,7 +52,15 @@ for astrophysical simulation codes.

.. toctree::
:maxdepth: 1
:caption: references
:caption: Unit tests

unit_tests
comprehensive_tests
one_zone_tests

.. toctree::
:maxdepth: 1
:caption: References

zreferences

Expand Down
103 changes: 103 additions & 0 deletions sphinx_docs/source/one_zone_tests.rst
Original file line number Diff line number Diff line change
@@ -0,0 +1,103 @@
**************
One Zone Tests
**************

There are several tests that let you call the EOS or reaction network
on a single zone to inspect the output directly.


``burn_cell``
=============

``burn_cell`` is a simple one-zone burn that will evolve a state with
a network for a specified amount of time. This can be used to
understand the timescales involved in a reaction sequence or to
determine the needed ODE tolerances.


Getting Started
---------------

The ``burn_cell`` code are located in
``Microphysics/unit_test/burn_cell``. To run a simulation, ensure that
both an input file and an initial conditions file have been created
and are in the same directory as the executable.

Input File
----------

These files are typically named as ``inputs_burn_network`` where network
is the network you wish to use for your testing.

The structure of this file is is fairly self-explanatory. The run
prefix defined should be unique to the tests that will be run as they
will be used to identify all of the output files. Typically, the run
prefix involves the name of the network being tested. The ``atol``
variables define absolute tolerances of the ordinary differential
equations and the ``rtol`` variables define the relative tolerances. The
second section of the input file collects the inputs that ``main.f90``
asks for so that the user does not have to input all 5+
parameters that are required every time the test is run. Each input
required is defined and initialized on the lines following
``&cellparams``. The use of the parameters is show below:

.. table:: The definition of parameters used in the burn_cell unit tests and specified in the second half of each inputs file.

+-----------------------+----------------------------------------+
| ``tmax`` | Maximum Time (s) |
+-----------------------+----------------------------------------+
| ``nsteps`` | Number of time subdivisions |
+-----------------------+----------------------------------------+
| ``density`` | State Density (:math:`\frac{g}{cm^3}`) |
+-----------------------+----------------------------------------+
| ``temperature`` | State Temperature (K) |
+-----------------------+----------------------------------------+
| ``massfractions(i)`` | Mass Fraction for element i |
+-----------------------+----------------------------------------+

Running the Code
----------------

To run the code, enter the burn_cell directory and run::

./main3d.gnu.ex inputs

where ``inputs`` is the name of your inputs file.

For each of the ``numsteps`` steps defined in the inputs
file, the code will output a files into a new directory titled
``run_prefix_output`` where ``run_prefix`` is the run prefix defined in the
inputs file. Each output file will be named using the run prefix
defined in the inputs file and the corresponding timestep.

Next, run ``burn_cell.py`` using python 3.x, giving the defined run prefix as an argument.
For example::

python3 burn_cell.py react_aprox13

The ``burn_cell.py`` code will gather information from all of the
output files and compile them into three graphs explained below.

Graphs Output by ``burn_cell.py``
---------------------------------

The file ``run-prefix_logX.png`` and ``run-prefix_logX.eps`` will display a
graph of the chemical abundances as a function of the time, both on
logarithmic scales, for all species involved in the simulation. An
example of this graph is shown below.

.. figure:: react_aprox13_logX.png
:alt: An example of a plot output by the burn_cell unit test. This is the logX output corresponding to the network aprox13.
:width: 4.5in

An example of a plot output by the burn_cell unit test. This is the
logX output corresponding to the network aprox13.



The file ``run-prefix_ydot.png`` and ``run-prefix_ydot.eps`` will display the
molar fraction (mass fraction / atomic weight) as a function of time,
both on logarithmic scales, for all species involved in the code.

The file ``run-prefix_T-edot.png`` and ``run-prefix_T-edot.eps`` will display
the temperature and the energy generation rate as a function of time.
Loading

0 comments on commit 8fd0d4a

Please sign in to comment.