With the recent changes, the coverage build time is down to a bit more than 2 hours which still feels quite long.
I therefore propose to increase the solver residual to 1e-3 and decrease Ls to 8 to speed up the test further. I expect a reduction in coverage build time by another ~60%.
I added a comment in the test which clarifies that one should not use such high residuals in physics workflows.
With the recent changes, the coverage build time is down to a bit more than 2 hours which still feels quite long. I therefore propose to increase the solver residual to 1e-3 and decrease Ls to 8 to speed up the test further. I expect a reduction in coverage build time by another ~60%.
I added a comment in the test which clarifies that one should not use such high residuals in physics workflows.