edef50f042
The support includes: - The standard fetch_add/fetch_sub operations. - Extra operations: (fetch_/opaque_)negate, (opaque_)add/sub. - Extra capability macros: BOOST_ATOMIC_FLOAT/DOUBLE/LONG_DOUBLE_LOCK_FREE. The atomic operations are currently implemented on top of the integer-based backends and thus are mostly CAS-based. The CAS operations perform binary comparisons, and as such have different behavior wrt. special FP values like NaN and signed zero than normal C++. The support for floating point types is optional and can be disabled by defining BOOST_ATOMIC_NO_FLOATING_POINT. This can be useful if on a certain platform parameters of the floating point types cannot be deduced from the compiler-defined or system macros (in which case the compilation fails). http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0020r6.html
79 lines
2.1 KiB
C++
79 lines
2.1 KiB
C++
// Copyright (c) 2018 Andrey Semashev
|
|
//
|
|
// Distributed under the Boost Software License, Version 1.0.
|
|
// See accompanying file LICENSE_1_0.txt or copy at
|
|
// http://www.boost.org/LICENSE_1_0.txt)
|
|
|
|
#ifndef BOOST_ATOMIC_TESTS_VALUE_WITH_EPSILON_H_INCLUDED_
|
|
#define BOOST_ATOMIC_TESTS_VALUE_WITH_EPSILON_H_INCLUDED_
|
|
|
|
#include <limits>
|
|
#include <iosfwd>
|
|
|
|
template< typename T >
|
|
class value_with_epsilon
|
|
{
|
|
private:
|
|
T m_value;
|
|
T m_epsilon;
|
|
|
|
public:
|
|
value_with_epsilon(T value, T epsilon) : m_value(value), m_epsilon(epsilon) {}
|
|
|
|
T value() const
|
|
{
|
|
return m_value;
|
|
}
|
|
|
|
T epsilon() const
|
|
{
|
|
return m_epsilon;
|
|
}
|
|
|
|
bool equal(T value) const
|
|
{
|
|
return value >= (m_value - m_epsilon) && value <= (m_value + m_epsilon);
|
|
}
|
|
|
|
friend bool operator== (T left, value_with_epsilon< T > const& right)
|
|
{
|
|
return right.equal(left);
|
|
}
|
|
friend bool operator== (value_with_epsilon< T > const& left, T right)
|
|
{
|
|
return left.equal(right);
|
|
}
|
|
|
|
friend bool operator!= (T left, value_with_epsilon< T > const& right)
|
|
{
|
|
return !right.equal(left);
|
|
}
|
|
friend bool operator!= (value_with_epsilon< T > const& left, T right)
|
|
{
|
|
return !left.equal(right);
|
|
}
|
|
};
|
|
|
|
template< typename Char, typename Traits, typename T >
|
|
inline std::basic_ostream< Char, Traits >& operator<< (std::basic_ostream< Char, Traits >& strm, value_with_epsilon< T > const& val)
|
|
{
|
|
// Note: libstdc++ does not provide output operators for __float128. There may also be no operators for long double.
|
|
// We don't use such floating point values in our tests where the cast would matter.
|
|
strm << static_cast< double >(val.value()) << " (+/-" << static_cast< double >(val.epsilon()) << ")";
|
|
return strm;
|
|
}
|
|
|
|
template< typename T, typename U >
|
|
inline value_with_epsilon< T > approx(T value, U epsilon)
|
|
{
|
|
return value_with_epsilon< T >(value, static_cast< T >(epsilon));
|
|
}
|
|
|
|
template< typename T >
|
|
inline value_with_epsilon< T > approx(T value)
|
|
{
|
|
return value_with_epsilon< T >(value, static_cast< T >(0.0000001));
|
|
}
|
|
|
|
#endif // BOOST_ATOMIC_TESTS_VALUE_WITH_EPSILON_H_INCLUDED_
|