Re:Can we trust floating-point numbers? (2006) [pdf]
by acqq on Tuesday September 16, 2014 @04:44PM
(#8327001)
You are absolutely right, 1e22 fits in the double, exactly with the bit pattern you showed. And it doesn't fit in the float. Thanks. One more proof that PDF alone is too little of the presentation and that it's easy for the reader to remain confused. I consider your reply "a missing slide" between p 15 and p 16.
So the author is: http://en.wikipedia.org/wiki/Paul_Zimmermann
Also 1e22 (note: not 10e22) is exactly representable in double (though not in float). Specifically, it is exactly equal to:
which requires only a 52 bit significand.