Objective:

Compare timing performance of matrix$-$matrix multiplication using floating point versus integer

arithmetic.

Task $1.$ Write c code to multiply two square matrixes of size NxN on LINUX platform. $($or MAC

in terminal window, use CLANG compiler$)$ using gcc$.$ For most inner loop that computes dot

product of row and column, create a function DP$($row$,$column$)$ and place it in separate file. You

have to compile and link two files main.c and dp$.$c$,$ and demonstrate that the new executable

runs.

For both integer and Floating$-$point arithmetic:

Taks $2.$ Measure execution time of matrix $x$ matrix multiplication for different values of $N($ at

least $5)$ using timing function from chrono library, as it was explained in class and reference was

posted on slack.

Task $3$ Create compiler generated assembly code for DP$,$ and compile DP in assembly, and link

with main to create executable. Run it and measure time as in Task $2.$

Task $4.$ Optimize compiler generated assembly code $($minimize the number of instructions, by

removing redundant instructions$).$ Perform timing measurements $s$ in Taks $2.$