Suppose we have an Consider the following C code to add a scalar s to a double type vector x[]: for (i=998; i>=0; i=i–1) x[i] = x[i] + s; Its MIPS code is: Loop: L. D F0,0(R1) #R1 is intital highest address of x[] ADD. D F4,F0,F2 #scalar s in F2 S. D F4,0(R1) #store result DADDUI R1,R1,-8 #decrease 8 bytes for DW BNE R1,R2,Loop #R2: address for stopping Please do Loop Unrolling and Rescheduling to optimize the MIPS code to reduce stalls by repeating every 3 iterations. What is the speed up after the Loop Unrolling and Rescheduling? application running on a 8-processor distributed-memory multiprocessor, which has a 100 ns time to handle reference to a remote memory. For this application, assume that all the references except those involving communication hit in the local memory hierarchy. Processors are stalled on a remote request, and the processor clock rate is 4 GHz. If the base CPI (assuming that all references hit in the cache) is 0.5, how much faster is the multiprocessor if there is no communication versus if 5% of the instructions involve a remote communication reference?