I am using the pow() function and I am trying to compare the return to a cpp_dec_float but I am getting an error
Code:
pow(sqrt(172.601), 2) != n);
Error:
UserPath\main.cpp:21: error: no match for 'operator!=' (operand types are '__gnu_cxx::__promote_2<double, int, double, double>::__type {aka double}' and 'boost::multiprecision::cpp_int {aka boost::multiprecision::number<boost::multiprecision::backends::cpp_int_backend<> >}')
pow(sqrt(172.601), 2) != n))
^
There are a number of pitfalls here. See the added links at the bottom.
Something tells me you've been using the backend type, not a frontend adaptor (like number<> or rational_adaptor<>).
The thing works without change:
#include <boost/multiprecision/cpp_dec_float.hpp>
#include <iostream>
int main() {
boost::multiprecision::cpp_dec_float_50 n = 3;
bool ok = pow(sqrt(172.601), 2) != n;
std::cout << std::boolalpha << ok;
}
Prints
true
You are mixing double and cpp_dec_float. This means that you do not gain much - if anything in the scenario - from the enhanced accuracy or decimal representation of Boost Multiprecision.
Instead consider going the whole way:
#include <boost/multiprecision/cpp_dec_float.hpp>
#include <iostream>
int main() {
typedef boost::multiprecision::cpp_dec_float_50 Decimal;
Decimal n("172.601");
Decimal other = pow(sqrt(Decimal("172.601")), 2);
std::cout << std::setprecision(50) << n << "\n";
std::cout << std::setprecision(50) << other << "\n";
bool equal = (abs(other - n) < std::numeric_limits<Decimal>::epsilon());
std::cout << std::boolalpha << equal;
}
Prints:
172.601
172.601
true
Note the CRUCIAL initialization from text, not a double literal!
Background info:
