/rust/registry/src/index.crates.io-1949cf8c6b5b557f/pxfm-0.1.29/src/bessel/y1f.rs
Line | Count | Source |
1 | | /* |
2 | | * // Copyright (c) Radzivon Bartoshyk 7/2025. All rights reserved. |
3 | | * // |
4 | | * // Redistribution and use in source and binary forms, with or without modification, |
5 | | * // are permitted provided that the following conditions are met: |
6 | | * // |
7 | | * // 1. Redistributions of source code must retain the above copyright notice, this |
8 | | * // list of conditions and the following disclaimer. |
9 | | * // |
10 | | * // 2. Redistributions in binary form must reproduce the above copyright notice, |
11 | | * // this list of conditions and the following disclaimer in the documentation |
12 | | * // and/or other materials provided with the distribution. |
13 | | * // |
14 | | * // 3. Neither the name of the copyright holder nor the names of its |
15 | | * // contributors may be used to endorse or promote products derived from |
16 | | * // this software without specific prior written permission. |
17 | | * // |
18 | | * // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
19 | | * // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
20 | | * // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
21 | | * // DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
22 | | * // FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
23 | | * // DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
24 | | * // SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
25 | | * // CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
26 | | * // OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
27 | | * // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
28 | | */ |
29 | | use crate::bessel::j0f::j1f_rsqrt; |
30 | | use crate::bessel::j1f::{j1f_asympt_alpha, j1f_asympt_beta}; |
31 | | use crate::bessel::trigo_bessel::cos_small; |
32 | | use crate::bessel::y1f_coeffs::{Y1_ZEROS, Y1_ZEROS_VALUES, Y1F_COEFFS}; |
33 | | use crate::common::f_fmla; |
34 | | use crate::double_double::DoubleDouble; |
35 | | use crate::logs::fast_logf; |
36 | | use crate::polyeval::{f_polyeval10, f_polyeval18, f_polyeval19}; |
37 | | use crate::rounding::CpuCeil; |
38 | | use crate::sincos_reduce::rem2pif_any; |
39 | | |
40 | | /// Bessel of the second kind of order 1 (Y1) |
41 | | /// |
42 | | /// Max ULP 0.5 |
43 | 0 | pub fn f_y1f(x: f32) -> f32 { |
44 | 0 | let ux = x.to_bits(); |
45 | 0 | if ux >= 0xffu32 << 23 || ux == 0 { |
46 | | // |x| == 0, |x| == inf, |x| == NaN, x < 0 |
47 | 0 | if ux.wrapping_shl(1) == 0 { |
48 | | // |x| == 0 |
49 | 0 | return f32::NEG_INFINITY; |
50 | 0 | } |
51 | | |
52 | 0 | if x.is_infinite() { |
53 | 0 | if x.is_sign_negative() { |
54 | 0 | return f32::NAN; |
55 | 0 | } |
56 | 0 | return 0.; |
57 | 0 | } |
58 | 0 | return x + f32::NAN; // x == NaN |
59 | 0 | } |
60 | | |
61 | 0 | let xb = x.to_bits(); |
62 | | |
63 | 0 | if xb <= 0x424e0000u32 { |
64 | | // x <= 51.5 |
65 | 0 | if xb <= 0x40000000u32 { |
66 | | // x <= 2 |
67 | 0 | if xb <= 0x3fb5c28fu32 { |
68 | | // x <= 1.42 |
69 | 0 | return y1f_near_zero(x); |
70 | 0 | } |
71 | | // transient zone from 1.42 to 2 have bad behavior for log poly already, |
72 | | // and not yet good to be easily covered, thus it use its own poly |
73 | 0 | return y1_transient_area(x); |
74 | 0 | } |
75 | 0 | return y1f_small_argument_path(x); |
76 | 0 | } |
77 | | |
78 | | // Exceptions |
79 | 0 | let bx = x.to_bits(); |
80 | 0 | if bx == 0x47037a3d { |
81 | 0 | return f32::from_bits(0x2deededb); |
82 | 0 | } else if bx == 0x65ce46e4 { |
83 | 0 | return f32::from_bits(0x9eed85c4); |
84 | 0 | } else if bx == 0x6bf68a7b { |
85 | 0 | return f32::from_bits(0x9dc70a09); |
86 | 0 | } else if bx == 0x76d84625 { |
87 | 0 | return f32::from_bits(0x15d7a68b); |
88 | 0 | } else if bx == 0x7e3dcda0 { |
89 | 0 | return f32::from_bits(0x12b81111); |
90 | 0 | } |
91 | | |
92 | 0 | y1f_asympt(x) |
93 | 0 | } |
94 | | |
95 | | /** |
96 | | Generated by SageMath: |
97 | | Evaluates: |
98 | | y2 = -J1(x)*log(x) + 1/x * (1 - sum((-1)^m*(H(m)+H(m-1))/(2^m*m!*(m-1)!)*x^(2*m)) |
99 | | Y1(x) = 2/pi*(-y2(x)+(euler_gamma - log(2))*J1(x)) |
100 | | expressed as: |
101 | | Y1(x)=log(x)*W1(x) - Z1(x) - 2/(pi*x) |
102 | | ```python |
103 | | from sage.all import * |
104 | | |
105 | | R = LaurentSeriesRing(RealField(300), 'x',default_prec=300) |
106 | | x = R.gen() |
107 | | N = 16 # Number of terms (adjust as needed) |
108 | | gamma = RealField(300)(euler_gamma) |
109 | | d2 = RealField(300)(2) |
110 | | pi = RealField(300).pi() |
111 | | log2 = RealField(300)(2).log() |
112 | | |
113 | | def j_series(n, x): |
114 | | return sum([(-1)**m * (x/2)**(ZZ(n) + ZZ(2)*ZZ(m)) / (ZZ(m).factorial() * (ZZ(m) + ZZ(n)).factorial()) for m in range(N)]) |
115 | | |
116 | | J1_series = j_series(1, x) |
117 | | |
118 | | def harmony(m): |
119 | | return sum(RealField(300)(1)/RealField(300)(k) for k in range(1, m+1)) |
120 | | |
121 | | def z_series(x): |
122 | | return sum([(-1)**m * (x)**(ZZ(2)*ZZ(m)) / (ZZ(2)**(2*m) * ZZ(m).factorial() * (ZZ(m) - ZZ(1)).factorial()) * (harmony(m) + harmony(m - 1)) for m in range(1, N)]) |
123 | | |
124 | | W1 = d2/pi * J1_series |
125 | | Z1 = -(d2/(x*pi) * z_series(x) + d2/pi * gamma * J1_series(x) - d2/pi * log2 * J1_series(x)) |
126 | | |
127 | | # see the series |
128 | | print(W0) |
129 | | print(Z0) |
130 | | ``` |
131 | | See ./notes/bessel_y1_taylor.ipynb for generation |
132 | | **/ |
133 | | #[inline] |
134 | 0 | fn y1f_near_zero(x: f32) -> f32 { |
135 | | const W: [u64; 10] = [ |
136 | | 0x3fd45f306dc9c883, |
137 | | 0xbfa45f306dc9c883, |
138 | | 0x3f5b2995e7b7b604, |
139 | | 0xbf021bb945252402, |
140 | | 0x3e9cf9286ea1d337, |
141 | | 0xbe2ee7a29824147f, |
142 | | 0x3db78be9987d036d, |
143 | | 0xbd3ae90af76a4d0f, |
144 | | 0x3cb7eb97f85e7d62, |
145 | | 0xbc31028e3376648a, |
146 | | ]; |
147 | 0 | let dx = x as f64; |
148 | 0 | let x2 = dx * dx; |
149 | 0 | let w0 = f_polyeval10( |
150 | 0 | x2, |
151 | 0 | f64::from_bits(W[0]), |
152 | 0 | f64::from_bits(W[1]), |
153 | 0 | f64::from_bits(W[2]), |
154 | 0 | f64::from_bits(W[3]), |
155 | 0 | f64::from_bits(W[4]), |
156 | 0 | f64::from_bits(W[5]), |
157 | 0 | f64::from_bits(W[6]), |
158 | 0 | f64::from_bits(W[7]), |
159 | 0 | f64::from_bits(W[8]), |
160 | 0 | f64::from_bits(W[9]), |
161 | 0 | ) * dx; |
162 | | const Z: [u64; 10] = [ |
163 | | 0x3fc91866143cbc8a, |
164 | | 0xbfabd3975c75b4a7, |
165 | | 0x3f6835b97894be5b, |
166 | | 0xbf12c7dbffcde97d, |
167 | | 0x3eb0a780ac776eac, |
168 | | 0xbe432e5a4ddeea30, |
169 | | 0x3dcf0ce34d2066a6, |
170 | | 0xbd52a4e1aea45c18, |
171 | | 0x3cd1474ade9154ac, |
172 | | 0xbc4978ba84f218c0, |
173 | | ]; |
174 | 0 | let z0 = f_polyeval10( |
175 | 0 | x2, |
176 | 0 | f64::from_bits(Z[0]), |
177 | 0 | f64::from_bits(Z[1]), |
178 | 0 | f64::from_bits(Z[2]), |
179 | 0 | f64::from_bits(Z[3]), |
180 | 0 | f64::from_bits(Z[4]), |
181 | 0 | f64::from_bits(Z[5]), |
182 | 0 | f64::from_bits(Z[6]), |
183 | 0 | f64::from_bits(Z[7]), |
184 | 0 | f64::from_bits(Z[8]), |
185 | 0 | f64::from_bits(Z[9]), |
186 | 0 | ) * dx; |
187 | 0 | let w_log = fast_logf(x); |
188 | | const TWO_OVER_PI: f64 = f64::from_bits(0x3fe45f306dc9c883); |
189 | 0 | let recip = 1. / dx; |
190 | 0 | let z = f_fmla(w0, w_log, -z0); |
191 | 0 | f_fmla(recip, -TWO_OVER_PI, z) as f32 |
192 | 0 | } |
193 | | |
194 | | #[inline] |
195 | 0 | fn y1_transient_area(x: f32) -> f32 { |
196 | 0 | let dx = x as f64; |
197 | | // first Y0 bessel zero |
198 | | const ZERO: DoubleDouble = |
199 | | DoubleDouble::from_bit_pair((0xbc8bd1e50d219bfd, 0x400193bed4dff243)); |
200 | 0 | let r = (dx - ZERO.hi) - ZERO.lo; |
201 | | /* |
202 | | Poly generated by Wolfram Matematica: |
203 | | <<FunctionApproximations` |
204 | | ClearAll["Global`*"] |
205 | | f[x_]:= BesselY[1,x + 2.1971413260310170351490335626990] |
206 | | {approx,error} = MiniMaxApproximation[f[x],{x,{1.42 - 2.1971413260310170351490335626990, 2 - 2.1971413260310170351490335626990 },17,0},WorkingPrecision->120] |
207 | | poly=error[[1]]; |
208 | | coeffs=CoefficientList[poly,x]; |
209 | | TableForm[Table[Row[{"'",NumberForm[coeffs[[i+1]],{50,50}, ExponentFunction->(Null&)],"',"}],{i,0,Length[coeffs]-1}]] |
210 | | */ |
211 | 0 | let p = f_polyeval18( |
212 | 0 | r, |
213 | 0 | f64::from_bits(0x3d9b15a8283b069b), |
214 | 0 | f64::from_bits(0x3fe0aa484455fd09), |
215 | 0 | f64::from_bits(0xbfbe56f80802fa38), |
216 | 0 | f64::from_bits(0xbfa0d2ac9d0409ad), |
217 | 0 | f64::from_bits(0xbf73a619b3551650), |
218 | 0 | f64::from_bits(0x3f7e6c480057ecbb), |
219 | 0 | f64::from_bits(0xbf650dc773a5df4d), |
220 | 0 | f64::from_bits(0x3f531e9ccab7d4da), |
221 | 0 | f64::from_bits(0xbf29b76999169b0e), |
222 | 0 | f64::from_bits(0x3f509c829abceaf7), |
223 | 0 | f64::from_bits(0x3f575aee5697c4d8), |
224 | 0 | f64::from_bits(0x3f63f7f9598be176), |
225 | 0 | f64::from_bits(0x3f67a6ae61541282), |
226 | 0 | f64::from_bits(0x3f665e6d3de19021), |
227 | 0 | f64::from_bits(0x3f5ee8837b9197f6), |
228 | 0 | f64::from_bits(0x3f4e6924f270fd7e), |
229 | 0 | f64::from_bits(0x3f32ca61e5b74925), |
230 | 0 | f64::from_bits(0x3f0725735bc3890b), |
231 | | ); |
232 | 0 | p as f32 |
233 | 0 | } |
234 | | |
235 | | /// This method on small range searches for nearest zero or extremum. |
236 | | /// Then picks stored series expansion at the point end evaluates the poly at the point. |
237 | | #[inline] |
238 | 0 | fn y1f_small_argument_path(x: f32) -> f32 { |
239 | 0 | let x_abs = x as f64; |
240 | | |
241 | | // let avg_step = 51.03 / 33.0; |
242 | | // let inv_step = 1.0 / avg_step; |
243 | | // |
244 | | // println!("inv_step {}", inv_step); |
245 | | |
246 | | const INV_STEP: f64 = 0.6466784244562023; |
247 | | |
248 | 0 | let fx = x_abs * INV_STEP; |
249 | | const Y1_ZEROS_COUNT: f64 = (Y1_ZEROS.len() - 1) as f64; |
250 | 0 | let idx0 = unsafe { fx.min(Y1_ZEROS_COUNT).to_int_unchecked::<usize>() }; |
251 | 0 | let idx1 = unsafe { |
252 | 0 | fx.cpu_ceil() |
253 | 0 | .min(Y1_ZEROS_COUNT) |
254 | 0 | .to_int_unchecked::<usize>() |
255 | | }; |
256 | | |
257 | 0 | let found_zero0 = DoubleDouble::from_bit_pair(Y1_ZEROS[idx0]); |
258 | 0 | let found_zero1 = DoubleDouble::from_bit_pair(Y1_ZEROS[idx1]); |
259 | | |
260 | 0 | let dist0 = (found_zero0.hi - x_abs).abs(); |
261 | 0 | let dist1 = (found_zero1.hi - x_abs).abs(); |
262 | | |
263 | 0 | let (found_zero, idx, dist) = if dist0 < dist1 { |
264 | 0 | (found_zero0, idx0, dist0) |
265 | | } else { |
266 | 0 | (found_zero1, idx1, dist1) |
267 | | }; |
268 | | |
269 | 0 | if idx == 0 { |
270 | | // Really should not happen here, but if it is then to log expansion |
271 | 0 | return y1f_near_zero(x); |
272 | 0 | } |
273 | | |
274 | | // We hit exact zero, value, better to return it directly |
275 | 0 | if dist == 0. { |
276 | 0 | return f64::from_bits(Y1_ZEROS_VALUES[idx]) as f32; |
277 | 0 | } |
278 | | |
279 | 0 | let c = &Y1F_COEFFS[idx - 1]; |
280 | | |
281 | 0 | let r = (x_abs - found_zero.hi) - found_zero.lo; |
282 | | |
283 | 0 | let p = f_polyeval19( |
284 | 0 | r, |
285 | 0 | f64::from_bits(c[0]), |
286 | 0 | f64::from_bits(c[1]), |
287 | 0 | f64::from_bits(c[2]), |
288 | 0 | f64::from_bits(c[3]), |
289 | 0 | f64::from_bits(c[4]), |
290 | 0 | f64::from_bits(c[5]), |
291 | 0 | f64::from_bits(c[6]), |
292 | 0 | f64::from_bits(c[7]), |
293 | 0 | f64::from_bits(c[8]), |
294 | 0 | f64::from_bits(c[9]), |
295 | 0 | f64::from_bits(c[10]), |
296 | 0 | f64::from_bits(c[11]), |
297 | 0 | f64::from_bits(c[12]), |
298 | 0 | f64::from_bits(c[13]), |
299 | 0 | f64::from_bits(c[14]), |
300 | 0 | f64::from_bits(c[15]), |
301 | 0 | f64::from_bits(c[16]), |
302 | 0 | f64::from_bits(c[17]), |
303 | 0 | f64::from_bits(c[18]), |
304 | | ); |
305 | | |
306 | 0 | p as f32 |
307 | 0 | } |
308 | | |
309 | | /* |
310 | | Evaluates: |
311 | | Y1 = sqrt(2/(PI*x)) * beta(x) * sin(x - 3*PI/4 - alpha(x)) |
312 | | |
313 | | Discarding 1/2*PI gives: |
314 | | Y1 = sqrt(2/(PI*x)) * beta(x) * (-cos(x - PI/4 - alpha(x))) |
315 | | */ |
316 | | #[inline] |
317 | 0 | fn y1f_asympt(x: f32) -> f32 { |
318 | 0 | let dx = x as f64; |
319 | | |
320 | 0 | let alpha = j1f_asympt_alpha(dx); |
321 | 0 | let beta = j1f_asympt_beta(dx); |
322 | | |
323 | 0 | let angle = rem2pif_any(x); |
324 | | |
325 | | const SQRT_2_OVER_PI: f64 = f64::from_bits(0x3fe9884533d43651); |
326 | | const MPI_OVER_4: f64 = f64::from_bits(0xbfe921fb54442d18); |
327 | | |
328 | 0 | let x0pi34 = MPI_OVER_4 - alpha; |
329 | 0 | let r0 = angle + x0pi34; |
330 | | |
331 | 0 | let m_cos = -cos_small(r0); |
332 | | |
333 | 0 | let z0 = beta * m_cos; |
334 | 0 | let scale = SQRT_2_OVER_PI * j1f_rsqrt(dx); |
335 | | |
336 | 0 | (scale * z0) as f32 |
337 | 0 | } |
338 | | |
339 | | #[cfg(test)] |
340 | | mod tests { |
341 | | use super::*; |
342 | | |
343 | | #[test] |
344 | | fn test_bessel_zero() { |
345 | | assert_eq!(f_y1f(700.76), 0.024876066); |
346 | | assert_eq!(f_y1f(35.76), 0.121432826); |
347 | | assert_eq!(f_y1f(1.76), -0.24787569); |
348 | | assert_eq!(f_y1f(0.87), -0.9030042); |
349 | | assert_eq!(f_y1f(f32::INFINITY), 0.0); |
350 | | assert!(f_y1f(f32::NEG_INFINITY).is_nan()); |
351 | | assert!(f_y1f(f32::NAN).is_nan()); |
352 | | } |
353 | | } |