Automatic Tests

If something changes in ProDiMo a suite of test models is automatically run with the newest ProDiMo version from the git repository (happens overnight). The results can be found here Automatic Test Results.

The idea of these automatic tests is mainly to find quickly problems caused by new code developments (i.e. compilation problems, new implementation causes other kinds of models to fail etc.). However, the automatic tests can also fail because a code change fixed a certain bug. This means that there was actually something wrong in the (physical) results of ProDiMo and we (the developers) now consider the new implementation as the correct one (or better one). In that case, the reference results for the automatic tests (the one used to compare with) are updated.

Below you will find a list of such code changes (starting with Jan 2020):

Log of significant code changes

This is a brief overview that summarizes code changes that might significantly affect the results of existing models (e.g. if one reruns an older model).

  • changeset:06fd19e7 (Jan 14, 2022) Fixed an issue with reading multiple T reactions from the KIDA database. This was a bug the fix can cause slight differences compared to older KIDA models. But seems to be insignificant for normal models.

  • changeset:519a78b4 (Dec 13, 2021) smallest grains are not settled any more. That should resolve inconsistencies in the outermost disk. This will for example change the dust to gas ratio in the outermost regions of the disk.

  • changeset:38955520 (Oct 14, 2021) a bugfix was made that should make the new chemistry solver more stable (for different compilers). Otherwise that same applies as for changeset:3644 (see below)

  • changeset:ad510383 (Sep 29, 2021) complex changed in the chemistry solver. This can cause some differences in the results but should affect only single points in the model. However, it might be that this can be different in more complex models.

  • changeset:fb126431 (Aug 13, 2021)

  • New default parameter for the radiative transfer to activate new code improvements (i.e. Diffusion Approximation). This should give more accurate results for the dust continuum RT. However, it will also affect the model results if one reruns an old model. To switch back to the old behaviour one can deactivate certain parameters (see RTparameter).

  • Also, the default value for Ninner was changed from 4 to 10. This will also affect the model results if Ninner was not set in the Parameter.in.

  • changeset:f5fd1142 (Jun 21, 2021) Dustbins are now available for the main branch. Adjustments to chemistry processes linked to dustbins. However, all changes should be backward compatible.

  • changeset:9d3ad824 (Sep 28, 2020) Several changes concerning the surface chemistry implementation. It is recommended to rerun older surface chemistry models with the newer ProDiMo version. Please use the updated Reactions.in : ReactionsDEV.in Results from models without surface chemistry should not be affected significantly.

  • changeset:b0dbadce (Sep 7, 2020) Add special desorption energy for PAHs in CR desorption. Bugfix, should not produce significant effects on model results (only CR PAH desorption rate changed).

  • changeset:e789fe73 (May 5, 2020): Implementation of position-dependent opacities. The contents of Mie.tmp has changed, now includes more variables needed to configure position-dependent opacities. "dust_opac.out" now gives opacity values based on dext_mono (~sca,~abs), instead of Qext_mono like before, and hence only for wavelengths that dext_mono is calculated.

  • changeset:4f1b284c (Jan 2, 2020): proper output of mass-averaged dust opacity across the whole disk. This changes the results of the dust opacities in dust_opac.out for models that for example use dust settling. However, this really only affects the output of the disk averaged dust opacities and not the radiative transfer or any other results of the model.

  • changeset:6ad8d407 (Dec 12, 2019): wrong collision rates for water. This can change the gas temperatures structure in the inner disk and affect line fluxes of mid-infrared lines.