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

Add complex32 and bcomplex32 #110

Open
hawkinsp opened this issue Oct 6, 2023 · 3 comments
Open

Add complex32 and bcomplex32 #110

hawkinsp opened this issue Oct 6, 2023 · 3 comments
Labels
enhancement New feature or request

Comments

@hawkinsp
Copy link
Collaborator

hawkinsp commented Oct 6, 2023

We'd need this for 16-bit FFTs in JAX: jax-ml/jax#17954

@hawkinsp hawkinsp added the enhancement New feature or request label Oct 6, 2023
@leofang
Copy link

leofang commented Jun 13, 2024

@hawkinsp @jakevdp @seberg @kmaehashi We already had a draft PR to add cupy.complex32 (cupy/cupy#4454) which is made as a legit numpy.dtype via NumPy C APIs (CuPy only uses NumPy dtypes; it does not have its own dtype system). My preference (speaking from NVIDIA perspective) is to add exotic NumPy dtypes to ml_dtypes and encourage all projects to use them. What would be the needed ingredients for me to migrate that PR over to here? Is there a guidance that I can follow along?

@jakevdp
Copy link
Collaborator

jakevdp commented Jun 13, 2024

I don't think there's any guidance beyond using the existing definitions in this repo as a template.

@matteoguarrera
Copy link

matteoguarrera commented Jun 27, 2024

Hey @leofang I am working on that as well! I can help. Dropping an email right now!
Pytorch does have complex32

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

No branches or pull requests

4 participants