/* * // Copyright (c) Radzivon Bartoshyk 6/2025. All rights reserved. * // * // Redistribution and use in source and binary forms, with or without modification, * // are permitted provided that the following conditions are met: * // * // 1. Redistributions of source code must retain the above copyright notice, this * // list of conditions and the following disclaimer. * // * // 2. Redistributions in binary form must reproduce the above copyright notice, * // this list of conditions and the following disclaimer in the documentation * // and/or other materials provided with the distribution. * // * // 3. Neither the name of the copyright holder nor the names of its * // contributors may be used to endorse or promote products derived from * // this software without specific prior written permission. * // * // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" * // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE * // DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE * // FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * // DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * // SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER * // CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, * // OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ use crate::common::*; use crate::double_double::DoubleDouble; use crate::sin::{ cos_accurate_near_zero, range_reduction_small, sin_accurate_near_zero, sincos_eval, }; use crate::sin_helper::sincos_eval_dd; use crate::sin_table::SIN_K_PI_OVER_128; use crate::sincos_reduce::LargeArgumentReduction; use std::hint::black_box; /// Sine and cosine for double precision /// /// ULP 0.5 pub fn f_sincos(x: f64) -> (f64, f64) { let x_e = (x.to_bits() >> 52) & 0x7ff; const E_BIAS: u64 = (1u64 << (11 - 1u64)) - 1u64; let y: DoubleDouble; let k; let mut argument_reduction = LargeArgumentReduction::default(); // |x| < 2^32 (with FMA) or |x| < 2^23 (w/o FMA) if x_e < E_BIAS + 16 { // |x| < 2^-26 let ax = x.to_bits() & 0x7fff_ffff_ffff_ffff; if ax <= 0x3fa921fbd34a9550 { // |x| <= 0.0490874 if x_e < E_BIAS - 27 { // Signed zeros. if x == 0.0 { return (x, 1.0); } // For |x| < 2^-26, |sin(x) - x| < ulp(x)/2. let s_sin = dyad_fmla(x, f64::from_bits(0xbc90000000000000), x); let s_cos = black_box(1.0) - min_normal_f64(); return (s_sin, s_cos); } // Polynomial for sin(x)/x // Generated by Sollya: // d = [0, 0.0490874]; // f_sin = sin(y)/y; // Q = fpminimax(f_sin, [|0, 2, 4, 6, 8|], [|1, D...|], d, relative, floating); const SIN_C: [u64; 4] = [ 0xbfc5555555555555, 0x3f8111111110e45a, 0xbf2a019ffd7fdaaf, 0x3ec71819b9bf01ef, ]; let x2 = x * x; let x4 = x2 * x2; let p01 = f_fmla(x2, f64::from_bits(SIN_C[1]), f64::from_bits(SIN_C[0])); let p23 = f_fmla(x2, f64::from_bits(SIN_C[3]), f64::from_bits(SIN_C[2])); let w0 = f_fmla(x4, p23, p01); let w1 = x2 * w0 * x; let r_sin = DoubleDouble::from_exact_add(x, w1); // Polynomial for cos(x) // Generated by Sollya: // d = [0, 0.0490874]; // f_cos = cos(y); // Q = fpminimax(f_cos, [|0, 2, 4, 6, 8|], [|1, D...|], d, relative, floating); const COS_C: [u64; 4] = [ 0xbfe0000000000000, 0x3fa55555555554a4, 0xbf56c16c1619b84a, 0x3efa013d3d01cf7f, ]; let p01 = f_fmla(x2, f64::from_bits(COS_C[1]), f64::from_bits(COS_C[0])); let p23 = f_fmla(x2, f64::from_bits(COS_C[3]), f64::from_bits(COS_C[2])); let w0 = f_fmla(x4, p23, p01); let w1 = x2 * w0; let r_cos = DoubleDouble::from_exact_add(1., w1); let err = f_fmla( x2, f64::from_bits(0x3cb0000000000000), // 2^-52 f64::from_bits(0x3be0000000000000), // 2^-65 ); let sin_ub = r_sin.hi + (r_sin.lo + err); let sin_lb = r_sin.hi + (r_sin.lo - err); let sin_x = if sin_ub == sin_lb { sin_ub } else { sin_accurate_near_zero(x) }; let cos_ub = r_cos.hi + (r_cos.lo + err); let cos_lb = r_cos.hi + (r_cos.lo - err); let cos_x = if cos_ub == cos_lb { cos_ub } else { cos_accurate_near_zero(x) }; return (sin_x, cos_x); } else { // // Small range reduction. (y, k) = range_reduction_small(x); } } else { // Inf or NaN if x_e > 2 * E_BIAS { // sin(+-Inf) = NaN return (x + f64::NAN, x + f64::NAN); } // Large range reduction. (k, y) = argument_reduction.reduce(x); } let r_sincos = sincos_eval(y); let (sin_y, cos_y) = (r_sincos.v_sin, r_sincos.v_cos); // Fast look up version, but needs 256-entry table. // cos(k * pi/128) = sin(k * pi/128 + pi/2) = sin((k + 64) * pi/128). let sk = SIN_K_PI_OVER_128[(k & 255) as usize]; let ck = SIN_K_PI_OVER_128[((k.wrapping_add(64)) & 255) as usize]; let sin_k = DoubleDouble::from_bit_pair(sk); let cos_k = DoubleDouble::from_bit_pair(ck); let msin_k = -sin_k; // After range reduction, k = round(x * 128 / pi) and y = x - k * (pi / 128). // So k is an integer and -pi / 256 <= y <= pi / 256. // Then sin(x) = sin((k * pi/128 + y) // = sin(y) * cos(k*pi/128) + cos(y) * sin(k*pi/128) let sin_k_cos_y = DoubleDouble::quick_mult(cos_y, sin_k); let cos_k_sin_y = DoubleDouble::quick_mult(sin_y, cos_k); // cos(x) = cos((k * pi/128 + y) // = cos(y) * cos(k*pi/128) - sin(y) * sin(k*pi/128) let cos_k_cos_y = DoubleDouble::quick_mult(cos_y, cos_k); let msin_k_sin_y = DoubleDouble::quick_mult(sin_y, msin_k); // cos_k_sin_y is always >> sin_k_cos_y let mut sin_dd = DoubleDouble::from_exact_add(sin_k_cos_y.hi, cos_k_sin_y.hi); // cos_k_cos_y is always >> msin_k_sin_y let mut cos_dd = DoubleDouble::from_exact_add(cos_k_cos_y.hi, msin_k_sin_y.hi); sin_dd.lo += sin_k_cos_y.lo + cos_k_sin_y.lo; cos_dd.lo += msin_k_sin_y.lo + cos_k_cos_y.lo; let sin_lp = sin_dd.lo + r_sincos.err; let sin_lm = sin_dd.lo - r_sincos.err; let cos_lp = cos_dd.lo + r_sincos.err; let cos_lm = cos_dd.lo - r_sincos.err; let sin_upper = sin_dd.hi + sin_lp; let sin_lower = sin_dd.hi + sin_lm; let cos_upper = cos_dd.hi + cos_lp; let cos_lower = cos_dd.hi + cos_lm; // Ziv's rounding test. if sin_upper == sin_lower && cos_upper == cos_lower { return (sin_upper, cos_upper); } sincos_hard(y, sin_k, cos_k, sin_upper, sin_lower, cos_upper, cos_lower) } #[cold] #[inline(never)] #[allow(clippy::too_many_arguments)] fn sincos_hard( y: DoubleDouble, sin_k: DoubleDouble, cos_k: DoubleDouble, sin_upper: f64, sin_lower: f64, cos_upper: f64, cos_lower: f64, ) -> (f64, f64) { let r_sincos = sincos_eval_dd(y); let msin_k = -sin_k; let sin_x = if sin_upper == sin_lower { sin_upper } else { // sin(x) = sin((k * pi/128 + u) // = sin(u) * cos(k*pi/128) + cos(u) * sin(k*pi/128) DoubleDouble::mul_add(sin_k, r_sincos.v_cos, cos_k * r_sincos.v_sin).to_f64() }; let cos_x = if cos_upper == cos_lower { cos_upper } else { // cos(x) = cos((k * pi/128 + u) // = cos(u) * cos(k*pi/128) - sin(u) * sin(k*pi/128) DoubleDouble::mul_add(cos_k, r_sincos.v_cos, msin_k * r_sincos.v_sin).to_f64() }; (sin_x, cos_x) } #[cfg(test)] mod tests { use super::*; #[test] fn f_sincos_test() { let subnormal = f_sincos(0.00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000015708065354637772); assert_eq!(subnormal.0, 1.5708065354637772e-307); assert_eq!(subnormal.1, 1.0); let zx_0 = f_sincos(0.0); assert_eq!(zx_0.0, 0.0); assert_eq!(zx_0.1, 1.0); let zx_1 = f_sincos(1.0); assert_eq!(zx_1.0, 0.8414709848078965); assert_eq!(zx_1.1, 0.5403023058681398); let zx_0_p5 = f_sincos(-0.5); assert_eq!(zx_0_p5.0, -0.479425538604203); assert_eq!(zx_0_p5.1, 0.8775825618903728); let zx_1 = f_sincos(0.002341235432); assert_eq!(zx_1.0, 0.0023412332931324344); assert_eq!(zx_1.1, 0.9999972593095778); let zx_1 = f_sincos(0.0198676543432); assert_eq!(zx_1.0, 0.019866347330026367); assert_eq!(zx_1.1, 0.9998026446473137); } }