Problem 4 (6 points) Consider the expression s1-3(4/3 1). When evaluated in double (1 point) Is this exactly, bitwise, the machine epsilon of double precision (eps in MATLAB)? (2 points) Consider a decimal floating point system (name it D5) with 5 decimal digits, precision, the result is 2.22044604925031e-16. Explain do.did2d3ds, and rounding to the nearest. Assume that s is evaluated in D5. Show all the intermediate steps in this evaluation (1 point) What is the machine epsilon in D5? How does the error in this evaluation compare to it? (2 points) Show the evaluation of s in D5 when the rounding is towards (a) -oo and (b) +oo. How do the corresponding errors compare to the machine epsilon of D5? Explain