Skip to content
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

Update CI to Julia v1.10 #29

Merged
merged 5 commits into from
Mar 13, 2024

Conversation

patrickersing
Copy link
Contributor

No description provided.

@patrickersing
Copy link
Contributor Author

I have adjusted the tolerances, such that all tests pass again. As the errors in the well-balanced tests are close to machine precision, I had to set the relative tolerance to rtol=10, while the absolute tolerance remains at a reasonable level atol=1e-10.
An alternative would be to reset the expected values, but then we will probably run into the same problem with the next version update.

@patrickersing patrickersing marked this pull request as ready for review March 11, 2024 13:09
@andrewwinters5000
Copy link
Member

andrewwinters5000 commented Mar 12, 2024

Wouldn't it be enough to only adjust the absolute tolerances on the well-balancing tests? The relative tolerances in the isapprox function do not hit a relative error. They are present to adjust for scaling. Instead, the function given two value x and y computes max( atol , rtol * max( abs(x), abs(y) ) ). For the well-balancedness tests the value of rtol * max( abs(x), abs(y) ) should be very small. I took the values from a failing test here to illustrate this:

julia> val1 = 4.998482888288039e-11
4.998482888288039e-11

julia> val2 = 4.704673109537872e-11
4.704673109537872e-11

julia> abs(val1-val2)
2.938097787501668e-12

julia> 1.4901161193847656e-8 * max( abs(val1), abs(val2) )
7.448319924306928e-19

julia> isapprox(val1, val2; atol = 1e-11)
true

Copy link
Member

@andrewwinters5000 andrewwinters5000 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! I would say we wait for #28 to get merged and then we can update the testing

@patrickersing
Copy link
Contributor Author

LGTM! I would say we wait for #28 to get merged and then we can update the testing

Sounds good!
Yes you are right, it works fine by just setting the atol. I falsely assumed isapprox was also comparing the relative error.

@andrewwinters5000 andrewwinters5000 enabled auto-merge (squash) March 13, 2024 09:56
@andrewwinters5000 andrewwinters5000 linked an issue Mar 13, 2024 that may be closed by this pull request
Copy link

codecov bot commented Mar 13, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.38%. Comparing base (4644c7e) to head (2b0d9fd).

Additional details and impacted files
@@            Coverage Diff             @@
##             main      #29      +/-   ##
==========================================
- Coverage   99.52%   99.38%   -0.14%     
==========================================
  Files          40       40              
  Lines        1459     1459              
==========================================
- Hits         1452     1450       -2     
- Misses          7        9       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@patrickersing
Copy link
Contributor Author

@andrewwinters5000 this should now be ready to merge

@andrewwinters5000 andrewwinters5000 enabled auto-merge (squash) March 13, 2024 11:11
@andrewwinters5000 andrewwinters5000 merged commit 50efb6d into trixi-framework:main Mar 13, 2024
7 of 9 checks passed
@patrickersing patrickersing deleted the ci_julia_1.10 branch March 13, 2024 12:01
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Test on Julia v1.10 (and possibly 1.9)
2 participants