GH-100425: Improve accuracy of builtin sum() for float inputs by rhettinger · Pull Request #100426 · python/cpython · GitHub

github.com

Tin mới

rhettinger:compensated_sum

GitHub CopilotWrite better code with AI

GitHub Copilot appDirect agents from issue to merge

ActionsAutomate any workflow

CodespacesInstant dev environments

IssuesPlan and track work

Code ReviewManage code changes

Code QualityEnforce quality at merge

GitHub Advanced SecurityFind and fix vulnerabilities

Code securitySecure your code as you build

Secret protectionStop leaks before they start

View all use cases

Financial services

View all industries

Software Development

GitHub SponsorsFund open source developers

Copilot for BusinessEnterprise-grade AI features

Premium SupportEnterprise-grade 24/7 support

Pull requests 2.6k

Security and quality 0

Improve accuracy of builtin sum() for float inputs #100425

Neumaier compensated summation

Revert "Kahan summation"

Restore the bogus sign tests

Add tests and comments

Add tests and docs

awaiting core review

Remove out of date example for fsum()

Update floating point tutorial

View reviewed changes

Learn more about hiding disruptive comments

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

Correct handling of overflowed or infinite sums

Move local declarations inside the block where they are used.

🔨 test-with-buildbots

For consistent results, apply compensation on the slow path as well.

Mark as an implementation detail.

Merged main into regmachine branch (

Update tutorial docs related to sum() accuracy

gh-100425: Update tutorial docs related to sum() accuracy #101854

Test failure in test_assert_bookmarks with Python 3.12 Kozea/WeasyPrint#1902

broken link to A.Neumaier article in built-in sum comment #111933

test python 3.12 pynucastro/pynucastro#710

Accuracy issues of sum() specialization for floats/complexes #122234

feat(retrieval): utterance-vs-knowledge document prior, inert at W=0 (#1174) robotrocketscience/aelfrice#1231

Add transparent NextGen rarity analysis tooling 6529-Collections/6529networkmuseum#4