169 8 bit floating point adder :: Quicker, easier and cheaper to make your own chip!

169 : 8 bit floating point adder

Author: Philip Mohr

Description: Adds two 8 Bit floating point numbers

Clock: 0 Hz

How it works

Adds two 8 bit floating point numbers under consideration of rounding and infinity cases. The two floats use the 8 bit input and the 8 bit bidirectional input. 1 bit sign, 4 bit exponent, 3 bit mantissa.

How to test

Every clock the output should give the addition of the two floats.

#	Input	Output	Bidirectional
0	Float a Mant[0]	Float out Mant[0]	Float b Mant[0]
1	Float a Mant[1]	Float out Mant[1]	Float b Mant[1]
2	Float a Mant[2]	Float out Mant[2]	Float b Mant[2]
3	Float a Exp[0]	Float out Exp[0]	Float b Exp[0]
4	Float a Exp[1]	Float out Exp[1]	Float b Exp[1]
5	Float a Exp[2]	Float out Exp[2]	Float b Exp[2]
6	Float a Exp[3]	Float out Exp[3]	Float b Exp[3]
7	Float a Sign	Float out Sign	Float b Sign

Input

Output

Bidirectional

Float a Mant[0]

Float out Mant[0]

Float b Mant[0]

Float a Mant[1]

Float out Mant[1]

Float b Mant[1]

Float a Mant[2]

Float out Mant[2]

Float b Mant[2]

Float a Exp[0]

Float out Exp[0]

Float b Exp[0]

Float a Exp[1]

Float out Exp[1]

Float b Exp[1]

Float a Exp[2]

Float out Exp[2]

Float b Exp[2]

Float a Exp[3]

Float out Exp[3]

Float b Exp[3]

Float a Sign

Float out Sign

Float b Sign

User feedback

MichaelBell: The adder works well for most values, but gives incorrect results for some small values. The linked script tests all possible values and logs results that might be incorrect. Link for more details

Chip location

169 8 bit floating point adder