-
Notifications
You must be signed in to change notification settings - Fork 99
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Different test values on julia v1.9.3 #1617
Comments
@andrewwinters5000 The failure occurs for one of the wet-dry elixirs. Could you maybe take a look at it and increase the test tolerance if required? |
I can try and carve out time soon to have a look. |
Thanks a lot! |
Investigating the results of # For julia 1.9.0
(l2 = [0.17979210479598923, 1.2377495706611434, 6.289818963361573e-8], linf = [0.845938394800688, 3.3740800777086575, 4.4541473087633676e-7])
# For julia 1.9.3
(l2 = [0.17979210239962704, 1.2377495649103047, 6.289758492103817e-8], linf = [0.8459383936345142, 3.374080078500849, 4.451690074347425e-7]) The fluid quantities pass but the bottom topography errors fail as, e.g., abs(l2_julia190[3] - l2_julia193[3]) = 6.047125775517298e-13 which is larger than the default We could adjust the tolerances, but I think a better way is to change this elixir to set # For julia 1.9.0
Final time: 10.0 Time steps: 15078 (accepted), 20324 (total)
# For julia 1.9.3
Final time: 10.0 Time steps: 15073 (accepted), 20387 (total) In contrast, when I run |
This is weird... Is there any discontinuous behavior involved in this setup that could explain these differences? |
What I meant was more some discontinuity in the discretization. Since this is a test case with drying, we do use some thresholds etc. that may explain this |
I think you could just use a slightly larger absolute tolerance in this case? |
That would be okay from my side, although it is interesting that the error based time stepping changes so noticeably. |
Yes, definitely... |
Recently, CI fails due to different test values in one elixir with a 1D
TreeMesh
, see e.g. here. I was able to reproduce the error locally on julia v1.9.3 (and v1.10.0-beta2), but not on julia v1.9.2.The text was updated successfully, but these errors were encountered: