The following is the representation of a 2-word floating point number:
|PDP-11 Floating Point Representations: (Word 1 of 2)|
|Sign Bit||Exponent||High-order fraction|
|PDP-11 Floating Point Representations: (Word 2 of 2)|
NOTE: The exponent is biased by 128.
For instance, 2.0 is 40200 in octal or 0100000010000000 in binary (remember the 128 bias and the hidden bit), plus a following 16-bit word of all zeros.
Four word floating point numbers add two more words for the
fraction of decreasing significance and increasing precision.