Files
xahaud/src/ripple/basics/impl/Number.cpp
Howard Hinnant 3f33471220 Introduce rounding modes for Number:
You can set a thread-local flag to direct Number how to round
non-exact results with the syntax:

    Number::rounding_mode prev_mode = Number::setround(Number::towards_zero);

This flag will stay in effect for this thread only until another call
to setround.  The previously set rounding mode is returned.

You can also retrieve the current rounding mode with:

    Number::rounding_mode current_mode = Number::getround();

The available rounding modes are:

* to_nearest : Rounds to nearest representable value.  On tie, rounds
               to even.
* towards_zero : Rounds towards zero.
* downward : Rounds towards negative infinity.
* upward : Rounds towards positive infinity.

The default rounding mode is to_nearest.
2023-02-07 15:43:28 -08:00

728 lines
17 KiB
C++
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

//------------------------------------------------------------------------------
/*
This file is part of rippled: https://github.com/ripple/rippled
Copyright (c) 2022 Ripple Labs Inc.
Permission to use, copy, modify, and/or distribute this software for any
purpose with or without fee is hereby granted, provided that the above
copyright notice and this permission notice appear in all copies.
THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
ANY SPECIAL , DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
//==============================================================================
#include <ripple/basics/Number.h>
#include <algorithm>
#include <cassert>
#include <numeric>
#include <stdexcept>
#include <type_traits>
#include <utility>
#ifdef _MSVC_LANG
#include <boost/multiprecision/cpp_int.hpp>
using uint128_t = boost::multiprecision::uint128_t;
#else // !defined(_MSVC_LANG)
using uint128_t = __uint128_t;
#endif // !defined(_MSVC_LANG)
namespace ripple {
thread_local Number::rounding_mode Number::mode_ = Number::to_nearest;
Number::rounding_mode
Number::getround()
{
return mode_;
}
Number::rounding_mode
Number::setround(rounding_mode mode)
{
return std::exchange(mode_, mode);
}
// Guard
// The Guard class is used to tempoarily add extra digits of
// preicision to an operation. This enables the final result
// to be correctly rounded to the internal precision of Number.
class Number::Guard
{
std::uint64_t digits_; // 16 decimal guard digits
std::uint8_t xbit_ : 1; // has a non-zero digit been shifted off the end
std::uint8_t sbit_ : 1; // the sign of the guard digits
public:
explicit Guard() : digits_{0}, xbit_{0}, sbit_{0}
{
}
// set & test the sign bit
void
set_positive() noexcept;
void
set_negative() noexcept;
bool
is_negative() const noexcept;
// add a digit
void
push(unsigned d) noexcept;
// recover a digit
unsigned
pop() noexcept;
// Indicate round direction: 1 is up, -1 is down, 0 is even
// This enables the client to round towards nearest, and on
// tie, round towards even.
int
round() noexcept;
};
inline void
Number::Guard::set_positive() noexcept
{
sbit_ = 0;
}
inline void
Number::Guard::set_negative() noexcept
{
sbit_ = 1;
}
inline bool
Number::Guard::is_negative() const noexcept
{
return sbit_ == 1;
}
inline void
Number::Guard::push(unsigned d) noexcept
{
xbit_ = xbit_ || (digits_ & 0x0000'0000'0000'000F) != 0;
digits_ >>= 4;
digits_ |= (d & 0x0000'0000'0000'000FULL) << 60;
}
inline unsigned
Number::Guard::pop() noexcept
{
unsigned d = (digits_ & 0xF000'0000'0000'0000) >> 60;
digits_ <<= 4;
return d;
}
// Returns:
// -1 if Guard is less than half
// 0 if Guard is exactly half
// 1 if Guard is greater than half
int
Number::Guard::round() noexcept
{
auto mode = Number::getround();
switch (mode)
{
case to_nearest:
if (digits_ > 0x5000'0000'0000'0000)
return 1;
if (digits_ < 0x5000'0000'0000'0000)
return -1;
if (xbit_)
return 1;
return 0;
case towards_zero:
return -1;
case downward:
if (sbit_)
{
if (digits_ > 0 || xbit_)
return 1;
}
return -1;
case upward:
if (sbit_)
return -1;
if (digits_ > 0 || xbit_)
return 1;
return -1;
}
}
// Number
constexpr Number one{1000000000000000, -15, Number::unchecked{}};
void
Number::normalize()
{
if (mantissa_ == 0)
{
*this = Number{};
return;
}
bool const negative = (mantissa_ < 0);
if (negative)
mantissa_ = -mantissa_;
auto m = static_cast<std::make_unsigned_t<rep>>(mantissa_);
while ((m < minMantissa) && (exponent_ > minExponent))
{
m *= 10;
--exponent_;
}
Guard g;
while (m > maxMantissa)
{
if (exponent_ >= maxExponent)
throw std::overflow_error("Number::normalize 1");
g.push(m % 10);
m /= 10;
++exponent_;
}
mantissa_ = m;
if ((exponent_ < minExponent) || (mantissa_ < minMantissa))
{
*this = Number{};
return;
}
auto r = g.round();
if (r == 1 || (r == 0 && (mantissa_ & 1) == 1))
{
++mantissa_;
if (mantissa_ > maxMantissa)
{
mantissa_ /= 10;
++exponent_;
}
}
if (exponent_ > maxExponent)
throw std::overflow_error("Number::normalize 2");
if (negative)
mantissa_ = -mantissa_;
}
Number&
Number::operator+=(Number const& y)
{
if (y == Number{})
return *this;
if (*this == Number{})
{
*this = y;
return *this;
}
if (*this == -y)
{
*this = Number{};
return *this;
}
assert(isnormal() && y.isnormal());
auto xm = mantissa();
auto xe = exponent();
int xn = 1;
if (xm < 0)
{
xm = -xm;
xn = -1;
}
auto ym = y.mantissa();
auto ye = y.exponent();
int yn = 1;
if (ym < 0)
{
ym = -ym;
yn = -1;
}
Guard g;
if (xe < ye)
{
if (xn == -1)
g.set_negative();
do
{
g.push(xm % 10);
xm /= 10;
++xe;
} while (xe < ye);
}
else if (xe > ye)
{
if (yn == -1)
g.set_negative();
do
{
g.push(ym % 10);
ym /= 10;
++ye;
} while (xe > ye);
}
if (xn == yn)
{
xm += ym;
if (xm > maxMantissa)
{
g.push(xm % 10);
xm /= 10;
++xe;
}
auto r = g.round();
if (r == 1 || (r == 0 && (xm & 1) == 1))
{
++xm;
if (xm > maxMantissa)
{
xm /= 10;
++xe;
}
}
if (xe > maxExponent)
throw std::overflow_error("Number::addition overflow");
}
else
{
if (xm > ym)
{
xm = xm - ym;
}
else
{
xm = ym - xm;
xe = ye;
xn = yn;
}
while (xm < minMantissa)
{
xm *= 10;
xm -= g.pop();
--xe;
}
auto r = g.round();
if (r == 1 || (r == 0 && (xm & 1) == 1))
{
--xm;
if (xm < minMantissa)
{
xm *= 10;
--xe;
}
}
if (xe < minExponent)
{
xm = 0;
xe = Number{}.exponent_;
}
}
mantissa_ = xm * xn;
exponent_ = xe;
return *this;
}
Number&
Number::operator*=(Number const& y)
{
if (*this == Number{})
return *this;
if (y == Number{})
{
*this = y;
return *this;
}
assert(isnormal() && y.isnormal());
auto xm = mantissa();
auto xe = exponent();
int xn = 1;
if (xm < 0)
{
xm = -xm;
xn = -1;
}
auto ym = y.mantissa();
auto ye = y.exponent();
int yn = 1;
if (ym < 0)
{
ym = -ym;
yn = -1;
}
auto zm = uint128_t(xm) * uint128_t(ym);
auto ze = xe + ye;
auto zn = xn * yn;
Guard g;
while (zm > maxMantissa)
{
g.push(static_cast<unsigned>(zm % 10));
zm /= 10;
++ze;
}
xm = static_cast<rep>(zm);
xe = ze;
auto r = g.round();
if (r == 1 || (r == 0 && (xm & 1) == 1))
{
++xm;
if (xm > maxMantissa)
{
xm /= 10;
++xe;
}
}
if (xe < minExponent)
{
xm = 0;
xe = Number{}.exponent_;
}
if (xe > maxExponent)
throw std::overflow_error(
"Number::multiplication overflow : exponent is " +
std::to_string(xe));
mantissa_ = xm * zn;
exponent_ = xe;
assert(isnormal() || *this == Number{});
return *this;
}
Number&
Number::operator/=(Number const& y)
{
if (y == Number{})
throw std::overflow_error("Number: divide by 0");
int np = 1;
auto nm = mantissa();
if (nm < 0)
{
nm = -nm;
np = -1;
}
int dp = 1;
auto dm = y.mantissa();
if (dm < 0)
{
dm = -dm;
dp = -1;
}
// Divide numerator and denominator such that the
// denominator is in the range [1, 10).
const int offset = -15 - y.exponent();
Number n{nm * (np * dp), exponent() + offset};
Number d{dm, y.exponent() + offset};
// Quadratic least squares fit to 1/x in the range [1, 10]
constexpr Number a0{9178756872006464, -16, unchecked{}};
constexpr Number a1{-2149215784206187, -16, unchecked{}};
constexpr Number a2{1405502114116773, -17, unchecked{}};
static_assert(a0.isnormal());
static_assert(a1.isnormal());
static_assert(a2.isnormal());
Number rm2{};
Number rm1{};
Number r = (a2 * d + a1) * d + a0;
// NewtonRaphson iteration of 1/x - d with initial guess r
// halt when r stops changing, checking for bouncing on the last iteration
do
{
rm2 = rm1;
rm1 = r;
r = r + r * (one - d * r);
} while (r != rm1 && r != rm2);
*this = n * r;
return *this;
}
Number::operator rep() const
{
rep drops = mantissa_;
int offset = exponent_;
Guard g;
if (drops != 0)
{
if (drops < 0)
{
g.set_negative();
drops = -drops;
}
for (; offset < 0; ++offset)
{
g.push(drops % 10);
drops /= 10;
}
for (; offset > 0; --offset)
{
if (drops > std::numeric_limits<decltype(drops)>::max() / 10)
throw std::overflow_error("Number::operator rep() overflow");
drops *= 10;
}
auto r = g.round();
if (r == 1 || (r == 0 && (drops & 1) == 1))
{
++drops;
}
if (g.is_negative())
drops = -drops;
}
return drops;
}
Number::operator XRPAmount() const
{
return XRPAmount{static_cast<rep>(*this)};
}
std::string
to_string(Number const& amount)
{
// keep full internal accuracy, but make more human friendly if possible
if (amount == Number{})
return "0";
auto const exponent = amount.exponent();
auto mantissa = amount.mantissa();
// Use scientific notation for exponents that are too small or too large
if (((exponent != 0) && ((exponent < -25) || (exponent > -5))))
{
std::string ret = std::to_string(mantissa);
ret.append(1, 'e');
ret.append(std::to_string(exponent));
return ret;
}
bool negative = false;
if (mantissa < 0)
{
mantissa = -mantissa;
negative = true;
}
assert(exponent + 43 > 0);
size_t const pad_prefix = 27;
size_t const pad_suffix = 23;
std::string const raw_value(std::to_string(mantissa));
std::string val;
val.reserve(raw_value.length() + pad_prefix + pad_suffix);
val.append(pad_prefix, '0');
val.append(raw_value);
val.append(pad_suffix, '0');
size_t const offset(exponent + 43);
auto pre_from(val.begin());
auto const pre_to(val.begin() + offset);
auto const post_from(val.begin() + offset);
auto post_to(val.end());
// Crop leading zeroes. Take advantage of the fact that there's always a
// fixed amount of leading zeroes and skip them.
if (std::distance(pre_from, pre_to) > pad_prefix)
pre_from += pad_prefix;
assert(post_to >= post_from);
pre_from = std::find_if(pre_from, pre_to, [](char c) { return c != '0'; });
// Crop trailing zeroes. Take advantage of the fact that there's always a
// fixed amount of trailing zeroes and skip them.
if (std::distance(post_from, post_to) > pad_suffix)
post_to -= pad_suffix;
assert(post_to >= post_from);
post_to = std::find_if(
std::make_reverse_iterator(post_to),
std::make_reverse_iterator(post_from),
[](char c) { return c != '0'; })
.base();
std::string ret;
if (negative)
ret.append(1, '-');
// Assemble the output:
if (pre_from == pre_to)
ret.append(1, '0');
else
ret.append(pre_from, pre_to);
if (post_to != post_from)
{
ret.append(1, '.');
ret.append(post_from, post_to);
}
return ret;
}
// Returns f^n
// Uses a log_2(n) number of multiplications
Number
power(Number const& f, unsigned n)
{
if (n == 0)
return one;
if (n == 1)
return f;
auto r = power(f, n / 2);
r *= r;
if (n % 2 != 0)
r *= f;
return r;
}
// Returns f^(1/d)
// Uses NewtonRaphson iterations until the result stops changing
// to find the non-negative root of the polynomial g(x) = x^d - f
// This function, and power(Number f, unsigned n, unsigned d)
// treat corner cases such as 0 roots as advised by Annex F of
// the C standard, which itself is consistent with the IEEE
// floating point standards.
Number
root(Number f, unsigned d)
{
if (f == one || d == 1)
return f;
if (d == 0)
{
if (f == -one)
return one;
if (abs(f) < one)
return Number{};
throw std::overflow_error("Number::root infinity");
}
if (f < Number{} && d % 2 == 0)
throw std::overflow_error("Number::root nan");
if (f == Number{})
return f;
// Scale f into the range (0, 1) such that f's exponent is a multiple of d
auto e = f.exponent() + 16;
auto const di = static_cast<int>(d);
auto ex = [e = e, di = di]() // Euclidean remainder of e/d
{
int k = (e >= 0 ? e : e - (di - 1)) / di;
int k2 = e - k * di;
if (k2 == 0)
return 0;
return di - k2;
}();
e += ex;
f = Number{f.mantissa(), f.exponent() - e}; // f /= 10^e;
bool neg = false;
if (f < Number{})
{
neg = true;
f = -f;
}
// Quadratic least squares curve fit of f^(1/d) in the range [0, 1]
auto const D = ((6 * di + 11) * di + 6) * di + 1;
auto const a0 = 3 * di * ((2 * di - 3) * di + 1);
auto const a1 = 24 * di * (2 * di - 1);
auto const a2 = -30 * (di - 1) * di;
Number r = ((Number{a2} * f + Number{a1}) * f + Number{a0}) / Number{D};
if (neg)
{
f = -f;
r = -r;
}
// NewtonRaphson iteration of f^(1/d) with initial guess r
// halt when r stops changing, checking for bouncing on the last iteration
Number rm1{};
Number rm2{};
do
{
rm2 = rm1;
rm1 = r;
r = (Number(d - 1) * r + f / power(r, d - 1)) / Number(d);
} while (r != rm1 && r != rm2);
// return r * 10^(e/d) to reverse scaling
return Number{r.mantissa(), r.exponent() + e / di};
}
Number
root2(Number f)
{
if (f == one)
return f;
if (f < Number{})
throw std::overflow_error("Number::root nan");
if (f == Number{})
return f;
// Scale f into the range (0, 1) such that f's exponent is a multiple of d
auto e = f.exponent() + 16;
if (e % 2 != 0)
++e;
f = Number{f.mantissa(), f.exponent() - e}; // f /= 10^e;
// Quadratic least squares curve fit of f^(1/d) in the range [0, 1]
auto const D = 105;
auto const a0 = 18;
auto const a1 = 144;
auto const a2 = -60;
Number r = ((Number{a2} * f + Number{a1}) * f + Number{a0}) / Number{D};
// NewtonRaphson iteration of f^(1/2) with initial guess r
// halt when r stops changing, checking for bouncing on the last iteration
Number rm1{};
Number rm2{};
do
{
rm2 = rm1;
rm1 = r;
r = (r + f / r) / Number(2);
} while (r != rm1 && r != rm2);
// return r * 10^(e/2) to reverse scaling
return Number{r.mantissa(), r.exponent() + e / 2};
}
// Returns f^(n/d)
Number
power(Number const& f, unsigned n, unsigned d)
{
if (f == one)
return f;
auto g = std::gcd(n, d);
if (g == 0)
throw std::overflow_error("Number::power nan");
if (d == 0)
{
if (f == -one)
return one;
if (abs(f) < one)
return Number{};
// abs(f) > one
throw std::overflow_error("Number::power infinity");
}
if (n == 0)
return one;
n /= g;
d /= g;
if ((n % 2) == 1 && (d % 2) == 0 && f < Number{})
throw std::overflow_error("Number::power nan");
return root(power(f, n), d);
}
} // namespace ripple