floating point

Discussion in 'General Electronics Chat' started by katherine1117, Dec 14, 2009.

  1. katherine1117

    Thread Starter New Member

    Oct 28, 2009
    when dealing with the 10-bit floating point arithmetic operations i find that when i added the exponents together i got one more bit than required,like: 0 1011 00010 and 0 1110 11001,then 1011+1110-0111(bias)=10010 but we only need 4 bits actually,how can i deal with it? thx
  2. kubeek

    AAC Fanatic!

    Sep 20, 2005
    I programming it would be called an arithmetic exception or overflow. Looks like you multiplied these numbers, so you added the exponents, and the resulting exponent is bigger than what can be represented by that 10bit model.