Machine precision is defined as the smallest machine number e such that 1+E > 1. Anything smaller, when added to 1, would be lost as roundoff. 3. Prove that e, as defined above, is the bound for relative round-off error. In other words, prove that b1-t chopping elbl-t rounding where b is the base of the computer's floating point number system and t is the mantissa length