From my research you can always fit it in a 32x32 multiply, you only need the extra bit at compile time. The extra bit tells you how to adjust the result in the

From my research you can always fit it in a 32x32 multiply, you only need the extra bit at compile time. The extra bit tells you how to adjust the result in the end, but the adjustment is also a constant.