You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We should add a radhydro problem to ensure that the output and performance does not change unexpctedly.
Describe alternatives you've considered
Leave regression tests as-is.
Additional context
We have many test problems with exact solutions for radiation, but we don't currently know if there are perturbations smaller than the error thresholds.
Describe the proposal
We currently only test hydro + optically-thin cooling in the regression test suite (https://quokka-astro.github.io/regression-testing-results/).
We should add a radhydro problem to ensure that the output and performance does not change unexpctedly.
Describe alternatives you've considered
Leave regression tests as-is.
Additional context
We have many test problems with exact solutions for radiation, but we don't currently know if there are perturbations smaller than the error thresholds.
We also don't know whether there are performance improvements/regressions over time. The regression test suite automatically tells us this, e.g.: https://quokka-astro.github.io/regression-testing-results/RandomBlast-GPU-timings.html
The text was updated successfully, but these errors were encountered: