-
Notifications
You must be signed in to change notification settings - Fork 125
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 InputError
to CUDA fft() function
#503
Comments
when going through Metal's
@xqft do you think I am in the right direction here. |
And to validate twiddle factors
we can do something like this please let me know if i missed something. |
Hello @startup-dreamer, sorry I just saw your comments! The logic of your first snippet is incorrect, FFT actually uses Your second snippet is doing the same as the first (you are defining Twiddle factors actually have some mathematical meaning, they're powers of a Nth primitive root of unity of the finite field which we are operating on, where You can checkout the |
We are removing this cuda implementation. See #831 |
The CPU and Metal
fft()
operation fns handle the case where an input with a non-power-of-two length is given and returns anInputError
in that case. CUDA's doesn't. This code should be removed andInputError
should be added to CUDA'sfft()
op.The text was updated successfully, but these errors were encountered: