rustfft/avx/
mod.rs

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
use crate::{Fft, FftDirection, FftNum};
use std::arch::x86_64::{__m256, __m256d};
use std::sync::Arc;

pub trait AvxNum: FftNum {
    type VectorType: AvxVector256<ScalarType = Self>;
}

impl AvxNum for f32 {
    type VectorType = __m256;
}
impl AvxNum for f64 {
    type VectorType = __m256d;
}

// Data that most (non-butterfly) SIMD FFT algorithms share
// Algorithms aren't required to use this struct, but it allows for a lot of reduction in code duplication
struct CommonSimdData<T, V> {
    inner_fft: Arc<dyn Fft<T>>,
    twiddles: Box<[V]>,

    len: usize,

    inplace_scratch_len: usize,
    outofplace_scratch_len: usize,

    direction: FftDirection,
}

macro_rules! boilerplate_avx_fft {
    ($struct_name:ident, $len_fn:expr, $inplace_scratch_len_fn:expr, $out_of_place_scratch_len_fn:expr) => {
        impl<A: AvxNum, T: FftNum> Fft<T> for $struct_name<A, T> {
            fn process_outofplace_with_scratch(
                &self,
                input: &mut [Complex<T>],
                output: &mut [Complex<T>],
                scratch: &mut [Complex<T>],
            ) {
                let required_scratch = self.get_outofplace_scratch_len();
                if scratch.len() < required_scratch
                    || input.len() < self.len()
                    || output.len() != input.len()
                {
                    // We want to trigger a panic, but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_outofplace(
                        self.len(),
                        input.len(),
                        output.len(),
                        self.get_outofplace_scratch_len(),
                        scratch.len(),
                    );
                    return; // Unreachable, because fft_error_outofplace asserts, but it helps codegen to put it here
                }

                let scratch = &mut scratch[..required_scratch];
                let result = array_utils::iter_chunks_zipped(
                    input,
                    output,
                    self.len(),
                    |in_chunk, out_chunk| {
                        self.perform_fft_out_of_place(in_chunk, out_chunk, scratch)
                    },
                );

                if result.is_err() {
                    // We want to trigger a panic, because the buffer sizes weren't cleanly divisible by the FFT size,
                    // but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_outofplace(
                        self.len(),
                        input.len(),
                        output.len(),
                        self.get_outofplace_scratch_len(),
                        scratch.len(),
                    )
                }
            }
            fn process_with_scratch(&self, buffer: &mut [Complex<T>], scratch: &mut [Complex<T>]) {
                let required_scratch = self.get_inplace_scratch_len();
                if scratch.len() < required_scratch || buffer.len() < self.len() {
                    // We want to trigger a panic, but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_inplace(
                        self.len(),
                        buffer.len(),
                        self.get_inplace_scratch_len(),
                        scratch.len(),
                    );
                    return; // Unreachable, because fft_error_inplace asserts, but it helps codegen to put it here
                }

                let scratch = &mut scratch[..required_scratch];
                let result = array_utils::iter_chunks(buffer, self.len(), |chunk| {
                    self.perform_fft_inplace(chunk, scratch)
                });

                if result.is_err() {
                    // We want to trigger a panic, because the buffer sizes weren't cleanly divisible by the FFT size,
                    // but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_inplace(
                        self.len(),
                        buffer.len(),
                        self.get_inplace_scratch_len(),
                        scratch.len(),
                    )
                }
            }
            #[inline(always)]
            fn get_inplace_scratch_len(&self) -> usize {
                $inplace_scratch_len_fn(self)
            }
            #[inline(always)]
            fn get_outofplace_scratch_len(&self) -> usize {
                $out_of_place_scratch_len_fn(self)
            }
        }
        impl<A: AvxNum, T> Length for $struct_name<A, T> {
            #[inline(always)]
            fn len(&self) -> usize {
                $len_fn(self)
            }
        }
        impl<A: AvxNum, T> Direction for $struct_name<A, T> {
            #[inline(always)]
            fn fft_direction(&self) -> FftDirection {
                self.direction
            }
        }
    };
}

macro_rules! boilerplate_avx_fft_commondata {
    ($struct_name:ident) => {
        impl<A: AvxNum, T: FftNum> Fft<T> for $struct_name<A, T> {
            fn process_outofplace_with_scratch(
                &self,
                input: &mut [Complex<T>],
                output: &mut [Complex<T>],
                scratch: &mut [Complex<T>],
            ) {
                if self.len() == 0 {
                    return;
                }

                let required_scratch = self.get_outofplace_scratch_len();
                if scratch.len() < required_scratch
                    || input.len() < self.len()
                    || output.len() != input.len()
                {
                    // We want to trigger a panic, but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_outofplace(
                        self.len(),
                        input.len(),
                        output.len(),
                        self.get_outofplace_scratch_len(),
                        scratch.len(),
                    );
                    return; // Unreachable, because fft_error_outofplace asserts, but it helps codegen to put it here
                }

                let scratch = &mut scratch[..required_scratch];
                let result = array_utils::iter_chunks_zipped(
                    input,
                    output,
                    self.len(),
                    |in_chunk, out_chunk| {
                        self.perform_fft_out_of_place(in_chunk, out_chunk, scratch)
                    },
                );

                if result.is_err() {
                    // We want to trigger a panic, because the buffer sizes weren't cleanly divisible by the FFT size,
                    // but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_outofplace(
                        self.len(),
                        input.len(),
                        output.len(),
                        self.get_outofplace_scratch_len(),
                        scratch.len(),
                    );
                }
            }
            fn process_with_scratch(&self, buffer: &mut [Complex<T>], scratch: &mut [Complex<T>]) {
                if self.len() == 0 {
                    return;
                }

                let required_scratch = self.get_inplace_scratch_len();
                if scratch.len() < required_scratch || buffer.len() < self.len() {
                    // We want to trigger a panic, but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_inplace(
                        self.len(),
                        buffer.len(),
                        self.get_inplace_scratch_len(),
                        scratch.len(),
                    );
                    return; // Unreachable, because fft_error_inplace asserts, but it helps codegen to put it here
                }

                let scratch = &mut scratch[..required_scratch];
                let result = array_utils::iter_chunks(buffer, self.len(), |chunk| {
                    self.perform_fft_inplace(chunk, scratch)
                });

                if result.is_err() {
                    // We want to trigger a panic, because the buffer sizes weren't cleanly divisible by the FFT size,
                    // but we want to avoid doing it in this function to reduce code size, so call a function marked cold and inline(never) that will do it for us
                    fft_error_inplace(
                        self.len(),
                        buffer.len(),
                        self.get_inplace_scratch_len(),
                        scratch.len(),
                    );
                }
            }
            #[inline(always)]
            fn get_inplace_scratch_len(&self) -> usize {
                self.common_data.inplace_scratch_len
            }
            #[inline(always)]
            fn get_outofplace_scratch_len(&self) -> usize {
                self.common_data.outofplace_scratch_len
            }
        }
        impl<A: AvxNum, T> Length for $struct_name<A, T> {
            #[inline(always)]
            fn len(&self) -> usize {
                self.common_data.len
            }
        }
        impl<A: AvxNum, T> Direction for $struct_name<A, T> {
            #[inline(always)]
            fn fft_direction(&self) -> FftDirection {
                self.common_data.direction
            }
        }
    };
}

#[macro_use]
mod avx_vector;

mod avx32_butterflies;
mod avx32_utils;

mod avx64_butterflies;
mod avx64_utils;

mod avx_bluesteins;
mod avx_mixed_radix;
mod avx_raders;

pub mod avx_planner;

pub use self::avx32_butterflies::{
    Butterfly11Avx, Butterfly128Avx, Butterfly12Avx, Butterfly16Avx, Butterfly24Avx,
    Butterfly256Avx, Butterfly27Avx, Butterfly32Avx, Butterfly36Avx, Butterfly48Avx,
    Butterfly512Avx, Butterfly54Avx, Butterfly5Avx, Butterfly64Avx, Butterfly72Avx, Butterfly7Avx,
    Butterfly8Avx, Butterfly9Avx,
};

pub use self::avx64_butterflies::{
    Butterfly11Avx64, Butterfly128Avx64, Butterfly12Avx64, Butterfly16Avx64, Butterfly18Avx64,
    Butterfly24Avx64, Butterfly256Avx64, Butterfly27Avx64, Butterfly32Avx64, Butterfly36Avx64,
    Butterfly512Avx64, Butterfly5Avx64, Butterfly64Avx64, Butterfly7Avx64, Butterfly8Avx64,
    Butterfly9Avx64,
};

pub use self::avx_bluesteins::BluesteinsAvx;
pub use self::avx_mixed_radix::{
    MixedRadix11xnAvx, MixedRadix12xnAvx, MixedRadix16xnAvx, MixedRadix2xnAvx, MixedRadix3xnAvx,
    MixedRadix4xnAvx, MixedRadix5xnAvx, MixedRadix6xnAvx, MixedRadix7xnAvx, MixedRadix8xnAvx,
    MixedRadix9xnAvx,
};
pub use self::avx_raders::RadersAvx2;
use self::avx_vector::AvxVector256;