Skip to content

Add 1 x 1 -> 2-limb SIMD convolution - #943

Draft
oscargus wants to merge 2 commits into
mainfrom
simdconv
Draft

Add 1 x 1 -> 2-limb SIMD convolution#943
oscargus wants to merge 2 commits into
mainfrom
simdconv

Conversation

@oscargus

@oscargus oscargus commented Apr 10, 2026

Copy link
Copy Markdown
Contributor

PR Summary

Closes #941

PR Checklist

  • "closes #0000" is in the body of the PR description to link the related issue
  • new and changed code is tested
  • relevant changes added to CHANGELOG.md
  • new functionality is documented

@codecov

codecov Bot commented Apr 10, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.68%. Comparing base (65ea944) to head (7e40ec2).

Additional details and impacted files
@@           Coverage Diff           @@
##             main     #943   +/-   ##
=======================================
  Coverage   98.67%   98.68%           
=======================================
  Files         125      125           
  Lines       36438    36497   +59     
  Branches     5042     5071   +29     
=======================================
+ Hits        35957    36016   +59     
+ Misses        427      425    -2     
- Partials       54       56    +2     
Flag Coverage Δ
hdl 98.93% <ø> (ø)
linux 84.54% <71.73%> (-0.03%) ⬇️
linux32 84.27% <76.92%> (-0.02%) ⬇️
macos 96.33% <100.00%> (+<0.01%) ⬆️
macos32 96.15% <100.00%> (+<0.01%) ⬆️
windows 98.06% <97.29%> (-0.01%) ⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@oscargus oscargus changed the title Add 32 x 32 -> 64-bit SIMD convolution Add 1 x 1 -> 2-limb SIMD convolution Apr 10, 2026
@oscargus
oscargus force-pushed the simdconv branch 2 times, most recently from 1fc7046 to 2d04a92 Compare April 10, 2026 16:11
@oscargus

Copy link
Copy Markdown
Contributor Author

This appears slower for the 64-bit case, at least for the 20-value case in the benchmark.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Use SIMD for 32 x 32 -> 64-bit inner products

1 participant