Get the high bits of the product of two overflowing integers. Like this, but more efficient for 128-bit integers:

function mul_hi(x::T, y::T) where T <: Base.BitInteger
    xy = widemul(x, y)
    (xy >> 8sizeof(T)) % T
end

mul_lo is just *.

mul_hi for BigInt could return 0, false, or error.

We could define mul_hi_lo instead to support non-overflowing integer types like SaferIntegers.

From discourse