Fix FPEs in tests #6265
Merged
Fix FPEs in tests #6265
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
wthrowe
requested changes
Sep 3, 2024
| T c(std::numeric_limits<T>::signaling_NaN()); | ||
| T d(std::numeric_limits<T>::signaling_NaN()); | ||
| // T c(std::numeric_limits<T>::quiet_NaN()); | ||
| // T d(std::numeric_limits<T>::quiet_NaN()); |
| @@ -350,13 +351,21 @@ bool FixConservatives::operator()( | |||
| : static_cast<SimdType>(0); | |||
|
|
|||
| // Equation B.27 of Foucart | |||
| // | |||
| // We avoid division by zero in the SIMD case using a mask and effectively | |||
| // set the result to zero by dividing by the max double. | |||
There was a problem hiding this comment.
It looks like you decided on a different implementation than this, without the max double part.
src/Time/ChooseLtsStepSize.cpp
Outdated
| @@ -22,7 +22,10 @@ TimeDelta choose_lts_step_size(const Time& time, const double desired_step) { | |||
| // n because of floating-point truncation. The inner ceil call | |||
| // avoids this problem. | |||
| const auto desired_step_power = | |||
| static_cast<size_t>(std::ceil(std::log2(std::ceil(desired_step_count)))); | |||
| desired_step == std::numeric_limits<double>::infinity() | |||
There was a problem hiding this comment.
This needs an abs(), since the step can be negative, or check for desired_step_count == 0.0 instead.
We should check for negative numbers before trying a c2p. This is also more efficient.
We need to get FPEs in tests enabled again so we can actually resolve these things before they get merged. I think getting these FPEs fixed will actually take some effort and the code isn't used right now anyway.
I don't know if this is the intended test, but it at least resolves the FPE. I also don't know why the quiet NaN causes an FPE on comparison or if the quiet NaN somehow gets changed to a signaling one in the middle.
nilsdeppe
force-pushed
the
fix_fpes_in_tests
branch
from
b00343e
to
bb24907
Compare
|
I made the changes. Since they are pretty small I just went ahead and squashed it. Partly to get this in before more FPEs appear 🙃 |
wthrowe
approved these changes
Sep 4, 2024
This was referenced Sep 11, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Proposed changes
There are a couple MC neutrino tests I commented out because they're too hard for me to fix. That code isn't currently being used in simulations and will need a lot more debugging and testing in general. I'll talk with Francois Foucart about getting that fixed separately. My main goal is to just get FPEs on again in tests.
Upgrade instructions
Code review checklist
make docto generate the documentation locally intoBUILD_DIR/docs/html.Then open
index.html.code review guide.
bugfixornew featureif appropriate.Further comments