How can I make sure that a variable is set to utilize eight-byte precision, in other words how can I achieve double precision for values stored in variables?
See the first link in
I would defer all division operations as late as possible in my calculations, and use integral values as long as possible. Integral values can grow surprisingly long.