Update overflow bound in log1p(f) #606
Merged
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
And add tests in this region. Fix functions in region between previous overflow bound and maximum float. Accuracy is maintained by relying on high accuracy log (in branch), however performance drops slightly. Ideally the core approximation should also work above threshold.
|
Note that performance drops only slightly due to introduction of branch, INFINITY case is now delegated to log. |
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Outdated suggestions cannot be applied.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.
Checklist
What is the purpose of this pull request?
What changes did you make?
This PR consists in fix overflow behaviour in log1p(f) scalar and SIMD.
Does this PR relate to any existing issue?
Fixes #473
Relates to #600 (where it is also reported)
Is there anything you would like reviewers to focus on?
Please on focus on method chosen to fix approximation between old oflow threshold and max float/double.
Interval was thoroughly tested for log1pf and log1p, but testing is always as good as it gets to double precision.