enzo-dev-problem-types / run /

Filename Size Date modified Message
3.1 KB
4.2 KB
1.6 KB
21.3 KB
The test_runner suite will run a series of tests with the current version of 
enzo.  It compares these new outputs against a set of existing "good" outputs 
(either from analytical solutions or from previous versions of enzo).  It 
returns how the current version of enzo performed on these tests.

You must specify a test destination directory outside of the enzo file 
structure where the tests will executed and tested.  You can also specify 
invididual test tags for choosing the tests to run.  By default, all the tests 
will be run.

In order to run the test_runner suite, go to the enzo/run subdirectory and 


Example: ./ -o ~/temp --AMR=True --author="John Wise"

The test tags are (with defaults first followed by alternatives):

name = ''
answer_testing_script = None  ''
nprocs = 1
runtime = 'short', 'medium', 'long'
hydro = False, True
mhd = False, True
gravity = False, True
cosmology = False, True
chemistry = False, True
cooling = False, True
AMR = False, True
dimensionality = 1, 2, 3
author = 'John Wise'
radiation = None, 'fld', 'ray'
max_time_minutes = 1  

Additional flags are:
  -h, --help            show this help message and exit
  -c COMPARE_DIR, --compare-dir=COMPARE_DIR
                        The directory structure to compare against
  --clobber             Recopies tests and tests from scratch.
  --interleave          Option to interleave preparation, running, and
  -m MACHINE, --machine=MACHINE
                        Machine to run tests on.
  -o OUTPUT_DIR, --output-dir=OUTPUT_DIR
                        Where to place the run directory
  --repo=REPOSITORY     Path to repository being tested.
  --sim-only            Only run simulations.
  --test-only           Only perform tests.
                        Multiply simulation time limit by this factor.
  -v, --verbose         Slightly more verbose output.

Three test suites exist: quick (37 tests), push (48 tests), and full (95 tests).  
The larger test suites encompass all tests of the smaller test suites.

The quick suite is designed to be run periodically as one works on the code.  The 
estimated run time is 25 minutes.
Run with --quicksuite=True
Gold standard available at:

The push suite should be run immediately prior to pushing changes or issuing a pull 
request.  The estimated run time is 90 minutes.
Run with --pullsuite=True
Gold standard available at:

The full suite is designed to be run monthly or whenever major changes have been pushed.  
The full suite contains tests that may take up to 10 hours to complete and also 
includes all the FLD tests which require the hypre-yes compile option and HYPRE libraries 
to be installed.  The estimated run time is about 36 hours.
Run with --fullsuite=True
Gold standard available at: