It's commonly used by everyone using "long double" instead of "double" in their x86 code.
long double isn't even supported by all compilers that run on x86-64 (most notably MSVC++), and of course it's specific to those processors. With other processors, long double can mean something else entirely.
Anyone serious about floating point only uses the IEEE754 formats, and for scientific computing the only one used is double precision, though single precision is also popular as a way to get an estimate on which you can apply iterative refinement.
For example, if you want to write a function that calculates sqrt (x^2 + y^2) for double precision x and y, that's very very hard to get right using double precision only, but using long double in the calculation and rounding to double makes it trivial.
Just use a library that does it correctly. The function is called 'hypot'.
Using long double just makes it slow and non-portable.