You are given the following code. Note that floating$-$point instructions use floating point registers labeled f$.$ Integer instructions use integer registers labeled x$.$

loop:

fld f$0,0($x$1)$

fld f$2,0($x$2)$

fadd f$6,$ f$2,$ f$12$

fsub f$4,$ f$0,$ f$10$

fmul f$8,$ f$4,$ f$6$

fsd f$8,0($x$2)$

subi x$1,$ x$1,8$

subi x$2,$ x$2,8$

bnz x$1,$ loop

We are given the following latencies for the different types of instructions.

Floating point Multiply: $5$

Floating Point Add$/$Sub:$3$

Integer arithmetic:$1$

Load$/$Store:$2$

Branch if not taken: $1$

Branch if taken: $2$

Remember how to use latencies. For example, since fadd and sub have latencies of $3,$ the fmul f$8,$ f$4,$ f$6$ which needs the results of fsub f$4,$ f$0,$ f$10$ must wait $3$ cycles from when fsub starts.

a$)$ Show how many cycles are needed to complete one iteration without reordering, but with data forwarding. DO NOT use delayed branch. Create a table that includes stalls $($not pipeline stages but stalls between instructions.

b$)$ Reorder instructions and show how many cycles are needed to complete one iteration. Show reordered code and cycles needed.

c$)$ Unroll the loop $2$ times, reorder instructions and show how many cycles are needed to complete two iterations.