Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CrossSIMD: Add a simple unit test, fix a couple of operations in the no-simd path #19955

Merged
merged 5 commits into from
Feb 6, 2025

Conversation

hrydgard
Copy link
Owner

@hrydgard hrydgard commented Feb 6, 2025

The no-SIMD path (used as fallback for RISC-V and other archs where CrossSIMD.h isn't implemented yet) still isn't quite working, but adding more of these tests will get us there.

@hrydgard hrydgard added this to the v1.19.0 milestone Feb 6, 2025
@hrydgard hrydgard merged commit d5bd7d2 into master Feb 6, 2025
19 checks passed
@hrydgard hrydgard deleted the crosssimd-test branch February 6, 2025 17:38
@fp64
Copy link
Contributor

fp64 commented Feb 8, 2025

Some comments (note: haven't actually compiled this, just looking at code).

Shouldn't Load4x3 use m[0..15], not just m[0..3]?
For Transpose doing it via local tmp buffer (so technically out-of-place) is probably faster.
AnyZeroSignBit technically wrong for -0.0f (probably mentioned this before).

@hrydgard
Copy link
Owner Author

hrydgard commented Feb 10, 2025

@fp64 Thanks for the catches and feedback, new PR upcoming (in which the depth rasterizer now works with the fallback).

(I'm not bothered by the AnyZeroSignBit one though, in practice it doesn't matter here)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants