Skip to content

feat: add SIMD float math functions (exp, exp2, log, log2, log10, sin… #400

New issue

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

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

Already on GitHub? # to your account

Merged
merged 1 commit into from
Mar 3, 2024

Conversation

avhz
Copy link

@avhz avhz commented Mar 3, 2024

Hi,

In relation to #14, I've added some math functions to check with you if I'm on the right track.

I simply followed the implementation that was already there for sqrt, and for functions that I easily found were already in intrinsics.

Let me know if it's the right way to go :)

Cheers!

Edit: I manually checked the output values against R, and they were all correct.

@calebzulawski
Copy link
Member

Thank you! I'm now noticing how lacking our tests are in this area, but we can leave that for another day.

@calebzulawski calebzulawski merged commit eea6f77 into rust-lang:master Mar 3, 2024
@avhz
Copy link
Author

avhz commented Mar 3, 2024

My pleasure :)

I did notice that too, and I didn't check any special values like T:MAX, infinity, etc.
But I can add tests for these.

In #14 you mentioned adding intrinsics to the compiler, is this completely done ?

@calebzulawski calebzulawski mentioned this pull request Mar 3, 2024
@calebzulawski
Copy link
Member

I actually thought the intrinsics you used here didn't exist yet. There might be some still missing, but none stand out to me right now. Also, I opened #401 to add the tests.

@calebzulawski
Copy link
Member

After a quick check, the intrinsics that seem to be missing are the pow/powi intrinsics

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

Successfully merging this pull request may close these issues.

2 participants