/* * // Copyright (c) Radzivon Bartoshyk 7/2025. 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::double_double::DoubleDouble; use crate::exponents::auxiliary::fast_ldexp; use crate::exponents::exp::{EXP_REDUCE_T0, EXP_REDUCE_T1, to_denormal}; use crate::exponents::expf::{ExpfBackend, GenericExpfBackend}; use crate::rounding::CpuRoundTiesEven; use std::hint::black_box; #[inline] fn exp10_poly_dd(z: DoubleDouble) -> DoubleDouble { const C: [(u64, u64); 6] = [ (0xbcaf48ad494ea102, 0x40026bb1bbb55516), (0xbcae2bfab318d399, 0x40053524c73cea69), (0x3ca81f50779e162b, 0x4000470591de2ca4), (0x3c931a5cc5d3d313, 0x3ff2bd7609fd98c4), (0x3c8910de8c68a0c2, 0x3fe1429ffd336aa3), (0xbc605e703d496537, 0x3fca7ed7086882b4), ]; let mut r = DoubleDouble::quick_mul_add( DoubleDouble::from_bit_pair(C[5]), z, DoubleDouble::from_bit_pair(C[4]), ); r = DoubleDouble::quick_mul_add(r, z, DoubleDouble::from_bit_pair(C[3])); r = DoubleDouble::quick_mul_add(r, z, DoubleDouble::from_bit_pair(C[2])); r = DoubleDouble::quick_mul_add(r, z, DoubleDouble::from_bit_pair(C[1])); DoubleDouble::quick_mul_add(r, z, DoubleDouble::from_bit_pair(C[0])) } #[cold] fn as_exp10_accurate(x: f64) -> f64 { let mut ix = x.to_bits(); let t = (f64::from_bits(0x40ca934f0979a371) * x).cpu_round_ties_even(); let jt: i64 = unsafe { t.to_int_unchecked::() // t is already integer, this is just a conversion }; let i1 = jt & 0x3f; let i0 = (jt >> 6) & 0x3f; let ie = jt >> 12; let t0 = DoubleDouble::from_bit_pair(EXP_REDUCE_T0[i0 as usize]); let t1 = DoubleDouble::from_bit_pair(EXP_REDUCE_T1[i1 as usize]); let dt = DoubleDouble::quick_mult(t0, t1); const L0: f64 = f64::from_bits(0x3f13441350800000); const L1: f64 = f64::from_bits(0xbd1f79fef311f12b); const L2: f64 = f64::from_bits(0xb9aac0b7c917826b); let dx = x - L0 * t; let dx_dd = DoubleDouble::quick_mult_f64(DoubleDouble::new(L2, L1), t); let dz = DoubleDouble::full_add_f64(dx_dd, dx); let mut f = exp10_poly_dd(dz); f = DoubleDouble::quick_mult(dz, f); let mut zfh: f64; if ix < 0xc0733a7146f72a42u64 { if (jt & 0xfff) == 0 { f = DoubleDouble::from_exact_add(f.hi, f.lo); let zt = DoubleDouble::from_exact_add(dt.hi, f.hi); f.hi = zt.lo; f = DoubleDouble::from_exact_add(f.hi, f.lo); ix = f.hi.to_bits(); if (ix.wrapping_shl(12)) == 0 { let l = f.lo.to_bits(); let sfh: i64 = ((ix as i64) >> 63) ^ ((l as i64) >> 63); ix = ix.wrapping_add(((1i64 << 51) ^ sfh) as u64); } zfh = zt.hi + f64::from_bits(ix); } else { f = DoubleDouble::quick_mult(f, dt); f = DoubleDouble::add(dt, f); f = DoubleDouble::from_exact_add(f.hi, f.lo); zfh = f.hi; } zfh = fast_ldexp(zfh, ie as i32); } else { ix = (1u64.wrapping_sub(ie as u64)) << 52; f = DoubleDouble::quick_mult(f, dt); f = DoubleDouble::add(dt, f); let zt = DoubleDouble::from_exact_add(f64::from_bits(ix), f.hi); f.hi = zt.hi; f.lo += zt.lo; zfh = to_denormal(f.to_f64()); } zfh } #[inline(always)] fn exp10_gen(x: f64, backend: B) -> f64 { let mut ix = x.to_bits(); let aix = ix & 0x7fff_ffff_ffff_ffff; if aix > 0x40734413509f79feu64 { // |x| > 0x40734413509f79fe if aix > 0x7ff0000000000000u64 { return x + x; } // nan if aix == 0x7ff0000000000000u64 { return if (ix >> 63) != 0 { 0.0 } else { x }; } if (ix >> 63) == 0 { return f64::from_bits(0x7fe0000000000000) * 2.0; // x > 308.255 } if aix > 0x407439b746e36b52u64 { // x < -323.607 return black_box(f64::from_bits(0x0018000000000000)) * black_box(f64::from_bits(0x3c80000000000000)); } } // check x integer to avoid a spurious inexact exception if ix.wrapping_shl(16) == 0 && (aix >> 48) <= 0x4036 { let kx = backend.round_ties_even(x); if kx == x { let k = kx as i64; if k >= 0 { let mut r = 1.0; for _ in 0..k { r *= 10.0; } return r; } } } /* avoid spurious underflow: for |x| <= 2.41082e-17 exp10(x) rounds to 1 to nearest */ if aix <= 0x3c7bcb7b1526e50eu64 { return 1.0 + x; // |x| <= 2.41082e-17 } let t = backend.round_ties_even(f64::from_bits(0x40ca934f0979a371) * x); let jt: i64 = unsafe { t.to_int_unchecked::() }; // t is already integer this is just a conversion let i1 = jt & 0x3f; let i0 = (jt >> 6) & 0x3f; let ie = jt >> 12; let t00 = EXP_REDUCE_T0[i0 as usize]; let t01 = EXP_REDUCE_T1[i1 as usize]; let t0 = DoubleDouble::from_bit_pair(t00); let t1 = DoubleDouble::from_bit_pair(t01); let mut tz = backend.quick_mult(t0, t1); const L0: f64 = f64::from_bits(0x3f13441350800000); const L1: f64 = f64::from_bits(0x3d1f79fef311f12b); let dx = backend.fma(-L1, t, backend.fma(-L0, t, x)); let dx2 = dx * dx; const CH: [u64; 4] = [ 0x40026bb1bbb55516, 0x40053524c73cea69, 0x4000470591fd74e1, 0x3ff2bd760a1f32a5, ]; let p0 = backend.fma(dx, f64::from_bits(CH[1]), f64::from_bits(CH[0])); let p1 = backend.fma(dx, f64::from_bits(CH[3]), f64::from_bits(CH[2])); let p = backend.fma(dx2, p1, p0); let mut fh = tz.hi; let fx = tz.hi * dx; let mut fl = backend.fma(fx, p, tz.lo); const EPS: f64 = 1.63e-19; if ix < 0xc0733a7146f72a42u64 { // x > -307.653 // x > -0x1.33a7146f72a42p+8 let ub = fh + (fl + EPS); let lb = fh + (fl - EPS); if lb != ub { return as_exp10_accurate(x); } fh = fast_ldexp(fh + fl, ie as i32); } else { // x <= -307.653: exp10(x) < 2^-1022 ix = 1u64.wrapping_sub(ie as u64).wrapping_shl(52); tz = DoubleDouble::from_exact_add(f64::from_bits(ix), fh); fl += tz.lo; let ub = fh + (fl + EPS); let lb = fh + (fl - EPS); if lb != ub { return as_exp10_accurate(x); } fh = to_denormal(fh + fl); } fh } #[cfg(any(target_arch = "x86", target_arch = "x86_64"))] #[target_feature(enable = "avx", enable = "fma")] unsafe fn exp10_fma_impl(x: f64) -> f64 { use crate::exponents::expf::FmaBackend; exp10_gen(x, FmaBackend {}) } /// Computes exp10 /// /// Max found ULP 0.5 pub fn f_exp10(x: f64) -> f64 { #[cfg(not(any(target_arch = "x86", target_arch = "x86_64")))] { exp10_gen(x, GenericExpfBackend {}) } #[cfg(any(target_arch = "x86", target_arch = "x86_64"))] { use std::sync::OnceLock; static EXECUTOR: OnceLock f64> = OnceLock::new(); let q = EXECUTOR.get_or_init(|| { if std::arch::is_x86_feature_detected!("avx") && std::arch::is_x86_feature_detected!("fma") { exp10_fma_impl } else { fn def_exp10(x: f64) -> f64 { exp10_gen(x, GenericExpfBackend {}) } def_exp10 } }); unsafe { q(x) } } } #[cfg(test)] mod tests { use super::*; #[test] fn test_exp10f() { assert_eq!(f_exp10(-3.370739843267434), 0.00042585343701025656); assert_eq!(f_exp10(1.), 10.0); assert_eq!(f_exp10(2.), 100.0); assert_eq!(f_exp10(3.), 1000.0); assert_eq!(f_exp10(4.), 10000.0); assert_eq!(f_exp10(5.), 100000.0); assert_eq!(f_exp10(6.), 1000000.0); assert_eq!(f_exp10(7.), 10000000.0); assert_eq!(f_exp10(f64::INFINITY), f64::INFINITY); assert_eq!(f_exp10(f64::NEG_INFINITY), 0.); assert!(f_exp10(f64::NAN).is_nan()); } }