Slide 20 of 67
Notes:
At the bottom of this diagram, we see something called the fractional bit. If we're multiplying two fractions, for instance a Q15 number times a Q15 number, then we want to have a Q31 output, so we have to shift the output of the multiplier one position to the left.
This is a zero overhead feature of the multiplier where by setting the fraction bit in the status register, we cause the multiplier to shift the output left by one bit, producing the desired result.
We call that a fraction bit. It's fairly common to have this sort of feature on a DSP, but it's a very easy feature to implement on the C54x.
Now if we look at the output of the multiplier, we see that it feeds into the adder...
