Ron Mak @ Apropos Logic

ron@apropos-logic.com
Publications and books by Ronald Mak

Java Number Cruncher

Program 3-2

IEEE 754 Floating-Point Formats

This applet displays the IEEE 754 single-precision and double-precision floating-point formats.

To run the demo:

  • The float and double panels are independent of each other.
  • Enter a floating-point number into the Value field and press the Decompose button.
  • Or, enter the individual parts of the number and press the Show value button.
  • Press the Enter key after typing a number to enter it into the field.
  • The Exponent bits, Biased value, and Unbiased value fields are interrelated. Changing one automatically changes the other two.
  • If you enter a reserved biased exponent number (0 and 255 for float, 0 and 2047 for double), the unbiased field will be disabled.
  • You do not need to enter the leading exponent 0 bits nor the trailing fraction 0 bits.
  • You can also enter into the Value field:
    • min, -min, max, -max for the minimum and maximum values of the type
    • inf or -inf for the infinity values of the type
    • nan for the Not-a-Number value of the type