feat(neon): add AArch32 compatibility for FMA intrinsics in neon_mathfun.h#6393
feat(neon): add AArch32 compatibility for FMA intrinsics in neon_mathfun.h#6393nihui merged 1 commit intoTencent:masterfrom
Conversation
|
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #6393 +/- ##
==========================================
- Coverage 95.89% 95.88% -0.01%
==========================================
Files 841 841
Lines 266338 266338
==========================================
- Hits 255402 255379 -23
- Misses 10936 10959 +23 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
The binary size change of libncnn.so (bytes)
|
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
|
Thanks for your contribution ! |
|
Thanks for your contribution ! |
The vfmaq_f32/vfmsq_f32 intrinsics are only available on AArch64.
To support ARM 32-bit (AArch32) targets, replace direct usage with
portable macros that fall back to vmlaq_f32/vmlsq_f32 on 32-bit NEON.
This enables successful compilation on armeabi-v7a while preserving
FMA performance on AArch64. All math functions (log, exp, sin, cos, etc.)
retain identical behavior and accuracy.