Closed Divisible8737 closed 3 years ago
Hi! Glad to hear you find the code useful.
The underlying library FINUFFT does indeed have single-precision support, but that has not been added to this interface yet. There's no timeplan available for it either, but contributions are of course welcome =)
Hey @y9KAaixxegjR , I have now implemented interfaces to the 32-bit calls of the library. Let me know if you encounter any issues with them.
First, thank you for this submission! The code works really, really well, and especially shines in 3D. This is a feature request:
Storing big 3D volumes as
ComplexF64
is very memory intensive. Does the core C++ code support complex floats instead of doubles? If so, wrapping theComplexF32
variant would be very useful for large 3D problems.