-
Notifications
You must be signed in to change notification settings - Fork 7
Open
Labels
enhancementNew feature or requestNew feature or request
Description
I think the major missing algorithm in this package is the prime-factorization algorithm for mixed-radix FFTs where the factors are co-prime. I'm not entirely sure, but I think this is why the composite benchmark is substantially worse than the rest relative to FFTW for small-size ffts.
This one has some serious considerations for a few reasons:
- which radices to choose in a many-composite algorithm?
- How to effectively incorporate this into the callgraph?
- Is this actually any better? I have no idea
For completeness, I should note that I've previously implemented Rader's algorithm at some point (I'm not sure if in C++ or julia...) and it did not work well at all unless your prime is length
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
enhancementNew feature or requestNew feature or request