The answers will vary on the following problems.  Please think and use the book
to complete the problem.  This is the type of problem that may appear on exams.
If you have doubts of your answer, work with someone to gain a better 
understanding.

Problem 6:

6.1] Write a code fragment for a simple loop to be executed on the DLX 5-stage
pipeline assuming full forwarding and bypassing, and separate instruction and
data memories.  The pipeline has FP ADD, FP MULT and FP DIV stages added to
the EX stage.  Include at least one instruction that utilizes one of these
floating point operations.  Make sure you record your assumptions and have 
a branch instruction at the end of the loop.

6.2] Compute the CPI of the loop you have written on the specified DLX machine.

6.3] Unroll the loop as many times as possible using rescheduling and renaming.
Act as a compiler and schedule an appropriate instruction to fill the branch 
delay slot.

6.4] Compute the CPI of the unrolled loop. Compare it to the first loop.
Did you effectively reduce stalls and control hazards? Did you reduce loop 
overhead?