Floating Point Types

Objectives

1. Discuss floats.

Notes

• float, double, long double
• All stored in IEEE 754 format
  • $+/-1.dddd x 2^{eeee}$
  • the number of d's and e's depend on the size.
  • Some $+/-0.00ddd x 2^{-abc}$
  • In this case, abc is determined by the format
    • -126 for floats
    • -16382 for long doubles
  • has inf, -inf and nan
• IO manip
  • reference
  • showpoint/noshowpont
  • fixed/scientific/defaultfloat
  • setprecision(n) (fixed first)
• casts
  • (type) express // c style
  • type(expression) // c++ style old
  • (type)(expression) // better
  • static_cast< type>(expression) new style
    • constant_cast
    • dynamic_cast
    • reinterpret_cast
• Don't forget cmath