35

Float arithmetic is hard

Comments
  • 6
    It is very hard, actually. CPUs have gotten a ton better at it though and a heck of a lot faster in the last decade!
  • 1
    I am hearing 'nonlinear dynamics' at university and see such results as pure evil since it may be exacly the point between stability or instabilility.
  • 1
    That's why we should never directly compare double or floats with == 😉
  • 2
    Floats are the jpeg of numbers: close, but imprecise and full of artifacts.
  • 0
    deconstruct and typecast to get consistency.
Add Comment