Skip to content

Revert "Use mul_add with a float inversion instead of mul_neg_add" #59

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.

Sign up for GitHub

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

Merged
Shnatsel merged 1 commit into from
Dec 13, 2025
Merged

Revert "Use mul_add with a float inversion instead of mul_neg_add" #59

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ num-traits = "0.2.19"
multiversion = "0.8.0"
num-complex = { version = "0.4.6", features = ["bytemuck"], optional = true }
bytemuck = { version = "1.23.2", optional = true }
wide = "1.0.3"
wide = "0.8.1"
rayon = { version = "1.11.0", optional = true }

[features]
Expand Down
37 changes: 20 additions & 17 deletions src/kernels/dit.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,7 +153,7 @@ pub fn fft_dit_chunk_8_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in1_im = f64x4::new(imags_s1[0..4].try_into().unwrap());

// out0.re = (in0.re + w.re * in1.re) - w.im * in1.im
let out0_re = sqrt2_2_im.mul_add(-in1_im, sqrt2_2.mul_add(in1_re, in0_re));
let out0_re = sqrt2_2_im.mul_neg_add(in1_im, sqrt2_2.mul_add(in1_re, in0_re));
// out0.im = (in0.im + w.re*in1.im) + w.im*in1.re
let out0_im = sqrt2_2_im.mul_add(in1_re, sqrt2_2.mul_add(in1_im, in0_im));

Expand Down Expand Up @@ -208,7 +208,7 @@ pub fn fft_dit_chunk_8_simd_f32(reals: &mut [f32], imags: &mut [f32]) {
let in1_im = f32x4::new(imags_s1[0..4].try_into().unwrap());

// out0.re = (in0.re + w.re * in1.re) - w.im * in1.im
let out0_re = sqrt2_2_im.mul_add(-in1_im, sqrt2_2.mul_add(in1_re, in0_re));
let out0_re = sqrt2_2_im.mul_neg_add(in1_im, sqrt2_2.mul_add(in1_re, in0_re));
// out0.im = (in0.im + w.re * in1.im) + w.im * in1.re
let out0_im = sqrt2_2_im.mul_add(in1_re, sqrt2_2.mul_add(in1_im, in0_im));

Expand Down Expand Up @@ -274,7 +274,7 @@ pub fn fft_dit_chunk_16_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in0_im = f64x8::new(imags_s0[0..8].try_into().unwrap());
let in1_im = f64x8::new(imags_s1[0..8].try_into().unwrap());

let out0_re = twiddle_im.mul_add(-in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_re = twiddle_im.mul_neg_add(in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_im = twiddle_im.mul_add(in1_re, twiddle_re.mul_add(in1_im, in0_im));

// out1 = 2*in0 - out0
Expand Down Expand Up @@ -339,7 +339,7 @@ pub fn fft_dit_chunk_16_simd_f32(reals: &mut [f32], imags: &mut [f32]) {
let in0_im = f32x8::new(imags_s0[0..8].try_into().unwrap());
let in1_im = f32x8::new(imags_s1[0..8].try_into().unwrap());

let out0_re = twiddle_im.mul_add(-in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_re = twiddle_im.mul_neg_add(in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_im = twiddle_im.mul_add(in1_re, twiddle_re.mul_add(in1_im, in0_im));

// out1 = 2*in0 - out0
Expand Down Expand Up @@ -426,8 +426,8 @@ pub fn fft_dit_chunk_32_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in0_im_0_7 = f64x8::new(imags_s0[0..8].try_into().unwrap());
let in1_im_0_7 = f64x8::new(imags_s1[0..8].try_into().unwrap());

let out0_re_0_7 =
twiddle_im_0_7.mul_add(-in1_im_0_7, twiddle_re_0_7.mul_add(in1_re_0_7, in0_re_0_7));
let out0_re_0_7 = twiddle_im_0_7
.mul_neg_add(in1_im_0_7, twiddle_re_0_7.mul_add(in1_re_0_7, in0_re_0_7));
let out0_im_0_7 =
twiddle_im_0_7.mul_add(in1_re_0_7, twiddle_re_0_7.mul_add(in1_im_0_7, in0_im_0_7));

Expand All @@ -445,8 +445,8 @@ pub fn fft_dit_chunk_32_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in0_im_8_15 = f64x8::new(imags_s0[8..16].try_into().unwrap());
let in1_im_8_15 = f64x8::new(imags_s1[8..16].try_into().unwrap());

let out0_re_8_15 = twiddle_im_8_15.mul_add(
-in1_im_8_15,
let out0_re_8_15 = twiddle_im_8_15.mul_neg_add(
in1_im_8_15,
twiddle_re_8_15.mul_add(in1_re_8_15, in0_re_8_15),
);
let out0_im_8_15 = twiddle_im_8_15.mul_add(
Expand Down Expand Up @@ -531,7 +531,7 @@ pub fn fft_dit_chunk_32_simd_f32(reals: &mut [f32], imags: &mut [f32]) {
let in0_im = f32x16::new(imags_s0[0..16].try_into().unwrap());
let in1_im = f32x16::new(imags_s1[0..16].try_into().unwrap());

let out0_re = twiddle_im.mul_add(-in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_re = twiddle_im.mul_neg_add(in1_im, twiddle_re.mul_add(in1_re, in0_re));
let out0_im = twiddle_im.mul_add(in1_re, twiddle_re.mul_add(in1_im, in0_im));

let out1_re = two.mul_sub(in0_re, out0_re);
Expand Down Expand Up @@ -665,7 +665,8 @@ pub fn fft_dit_chunk_64_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in0_im = f64x8::new(imags_s0[0..8].try_into().unwrap());
let in1_im = f64x8::new(imags_s1[0..8].try_into().unwrap());

let out0_re = twiddle_im_0_7.mul_add(-in1_im, twiddle_re_0_7.mul_add(in1_re, in0_re));
let out0_re =
twiddle_im_0_7.mul_neg_add(in1_im, twiddle_re_0_7.mul_add(in1_re, in0_re));
let out0_im = twiddle_im_0_7.mul_add(in1_re, twiddle_re_0_7.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
let out1_im = two.mul_sub(in0_im, out0_im);
Expand All @@ -681,7 +682,8 @@ pub fn fft_dit_chunk_64_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in0_im = f64x8::new(imags_s0[8..16].try_into().unwrap());
let in1_im = f64x8::new(imags_s1[8..16].try_into().unwrap());

let out0_re = twiddle_im_8_15.mul_add(-in1_im, twiddle_re_8_15.mul_add(in1_re, in0_re));
let out0_re =
twiddle_im_8_15.mul_neg_add(in1_im, twiddle_re_8_15.mul_add(in1_re, in0_re));
let out0_im = twiddle_im_8_15.mul_add(in1_re, twiddle_re_8_15.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
let out1_im = two.mul_sub(in0_im, out0_im);
Expand All @@ -698,7 +700,7 @@ pub fn fft_dit_chunk_64_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in1_im = f64x8::new(imags_s1[16..24].try_into().unwrap());

let out0_re =
twiddle_im_16_23.mul_add(-in1_im, twiddle_re_16_23.mul_add(in1_re, in0_re));
twiddle_im_16_23.mul_neg_add(in1_im, twiddle_re_16_23.mul_add(in1_re, in0_re));
let out0_im =
twiddle_im_16_23.mul_add(in1_re, twiddle_re_16_23.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
Expand All @@ -716,7 +718,7 @@ pub fn fft_dit_chunk_64_simd_f64(reals: &mut [f64], imags: &mut [f64]) {
let in1_im = f64x8::new(imags_s1[24..32].try_into().unwrap());

let out0_re =
twiddle_im_24_31.mul_add(-in1_im, twiddle_re_24_31.mul_add(in1_re, in0_re));
twiddle_im_24_31.mul_neg_add(in1_im, twiddle_re_24_31.mul_add(in1_re, in0_re));
let out0_im =
twiddle_im_24_31.mul_add(in1_re, twiddle_re_24_31.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
Expand Down Expand Up @@ -836,7 +838,8 @@ pub fn fft_dit_chunk_64_simd_f32(reals: &mut [f32], imags: &mut [f32]) {
let in0_im = f32x16::new(imags_s0[0..16].try_into().unwrap());
let in1_im = f32x16::new(imags_s1[0..16].try_into().unwrap());

let out0_re = twiddle_im_0_15.mul_add(-in1_im, twiddle_re_0_15.mul_add(in1_re, in0_re));
let out0_re =
twiddle_im_0_15.mul_neg_add(in1_im, twiddle_re_0_15.mul_add(in1_re, in0_re));
let out0_im = twiddle_im_0_15.mul_add(in1_re, twiddle_re_0_15.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
let out1_im = two.mul_sub(in0_im, out0_im);
Expand All @@ -853,7 +856,7 @@ pub fn fft_dit_chunk_64_simd_f32(reals: &mut [f32], imags: &mut [f32]) {
let in1_im = f32x16::new(imags_s1[16..32].try_into().unwrap());

let out0_re =
twiddle_im_16_31.mul_add(-in1_im, twiddle_re_16_31.mul_add(in1_re, in0_re));
twiddle_im_16_31.mul_neg_add(in1_im, twiddle_re_16_31.mul_add(in1_re, in0_re));
let out0_im =
twiddle_im_16_31.mul_add(in1_re, twiddle_re_16_31.mul_add(in1_im, in0_im));
let out1_re = two.mul_sub(in0_re, out0_re);
Expand Down Expand Up @@ -911,7 +914,7 @@ pub fn fft_dit_64_chunk_n_simd(
let tw_im = f64x8::new(*tw_im);

// out0.re = (in0.re + tw_re * in1.re) - tw_im * in1.im
let out0_re = tw_im.mul_add(-in1_im, tw_re.mul_add(in1_re, in0_re));
let out0_re = tw_im.mul_neg_add(in1_im, tw_re.mul_add(in1_re, in0_re));
// out0.im = (in0.im + tw_re * in1.im) + tw_im * in1.re
let out0_im = tw_im.mul_add(in1_re, tw_re.mul_add(in1_im, in0_im));

Expand Down Expand Up @@ -972,7 +975,7 @@ pub fn fft_dit_32_chunk_n_simd(
let tw_im = f32x16::new(*tw_im);

// out0.re = (in0.re + tw_re * in1.re) - tw_im * in1.im
let out0_re = tw_im.mul_add(-in1_im, tw_re.mul_add(in1_re, in0_re));
let out0_re = tw_im.mul_neg_add(in1_im, tw_re.mul_add(in1_re, in0_re));
// out0.im = (in0.im + tw_re * in1.im) + tw_im * in1.re
let out0_im = tw_im.mul_add(in1_re, tw_re.mul_add(in1_im, in0_im));

Expand Down
Loading