Why do we use mid = low + (high – low)/2; but not mid = (low/2) +(high /2)?

Question

In binary search, we use mid = low + (high – low)/2 instead of (low + high)/2 to avoid overflow, however, can't calculate low/2 and high/2 separately and then sum them up rather than low+(( high-low)/2)?

P.S. If low + (high – low)/2 is more efficient, then why is it so?

Answer 1

Can't we calculate low/2 and high/2 separately and then sum them up rather than using low+((high-low)/2)?

Sure.

If low+(high-low)/2 is more efficient, then why is it so?

For a lot of hardware dividing is slower than adding and subtracting, so dividing twice might be slower than the method that uses more adding and subtracting.