Struct ScaledDitButterfly 

pub struct ScaledDitButterfly<F> {
    pub twiddle: F,
    pub scale: F,
    pub twiddle_times_scale: F,
}

DIT (Decimation-In-Time) butterfly operation with a post-multiplication scale factor.

This butterfly computes:

  output_1 = (x1 + x2 * twiddle) * scale
  output_2 = (x1 - x2 * twiddle) * scale

which is equivalent to:

  output_1 = x1 * scale + x2 * (twiddle * scale)
  output_2 = x1 * scale - x2 * (twiddle * scale)

This is used to merge a uniform scaling step (e.g., 1/N normalization in inverse DFT) into a butterfly pass, avoiding a separate memory pass over the data.

The struct stores scale and twiddle_times_scale = twiddle * scale so that the apply method only needs 2 multiplications instead of 3.

Fields

twiddle: F

scale: F

twiddle_times_scale: F

Precomputed product twiddle * scale to reduce multiplications in the hot loop.

Implementations

impl<F: Field> ScaledDitButterfly<F>

pub fn new(twiddle: F, scale: F) -> Self

Construct a ScaledDitButterfly, precomputing twiddle * scale.

Trait Implementations

impl<F: Field> Butterfly<F> for ScaledDitButterfly<F>

fn apply_to_rows(&self, row_1: &mut [F], row_2: &mut [F])

Override apply_to_rows to pre-broadcast both scale and twiddle_times_scale into packed fields once before the inner loop.

fn apply<PF: PackedField<Scalar = F>>(&self, x_1: PF, x_2: PF) -> (PF, PF)

Applies the butterfly transformation to two packed field values.

fn apply_in_place<PF: PackedField<Scalar = F>>( &self, x_1: &mut PF, x_2: &mut PF, )

Applies the butterfly in-place to two packed values.

fn apply_to_rows_oop( &self, src_1: &[F], dst_1: &mut [MaybeUninit<F>], src_2: &[F], dst_2: &mut [MaybeUninit<F>], )

Applies the butterfly out-of-place to two source rows.

impl<F: Clone> Clone for ScaledDitButterfly<F>

fn clone(&self) -> ScaledDitButterfly<F>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<F: Copy> Copy for ScaledDitButterfly<F>