Precision long double worse than double

Question

When I use long double I get worse precision than using double. 3.14159265358979323846264L is this good to write long double const in source code or I should add something other than L?

EDIT I solved the problem. I change algorithm to be more precisely.

Answer 1

Your are not getting a worse precision.

What is happening is that when you print the number out the stream library is truncating the displayed value. Use std::setprecision to get a specific precision.

double        x = 1.0/3;
long double   y = 1.0/6;

// Prints out the precision
std::cout << "Limit: " << std::numeric_limits<double>::digits10 << "\n";
std::cout << "Limit: " << std::numeric_limits<long double>::digits10 << "\n";

// prints the numbers with max precision.
std::cout << std::setprecision(std::numeric_limits<double>::digits10) << x << "\n";
std::cout << std::setprecision(std::numeric_limits<long double>::digits10) << y << "\n";