tests: test-progs: matrix-mult: Add matrix multiplication code for testing

Introduce matrix multiplication test code which allows us to test
performance of various cache replacement algorithms.

This needs to compiled as follows
g++ -static-libgcc mm.cpp -O2 -o mm -lm libpthread.a m5op_x86.o

The pthread static library for x86 is available from this branch
https://github.com/tiwarianoop2/m5threads

and m5op_x86 needs to be generated as described here
http://pages.cs.wisc.edu/~david/courses/cs752/Fall2015/wiki/index.php?n=Main.Homework3

This has been tested only for x86.

tests/test-progs/matrix-mult/mm.cpp

@@ -0,0 +1,82 @@
+#include <iomanip>
+#include <iostream>
+#include <random>
+
+#include "../../../util/m5/m5op.h"
+
+#define BLOCK_SIZE 16
+
+using namespace std;
+
+void multiply(double **A, double **B, double **C, int size)
+{
+    for (int k = 0; k < size; k += BLOCK_SIZE) {
+        for (int j = 0; j < size; j += BLOCK_SIZE) {
+            for (int i = 0; i < size; i++) {
+                for (int jj = j; jj < min(j + BLOCK_SIZE, size); jj++) {
+                    for (int kk = k; kk < min(k + BLOCK_SIZE, size); kk++) {
+                        C[i][jj] += A[i][kk] * B[kk][jj];
+                    }
+                }
+            }
+        }
+    }
+}
+
+void printMatrix(double **A, int size)
+{
+    for (int i = 0; i < size; i++) {
+        for (int j = 0; j < size; j++) {
+            cout << setprecision(3) << setw(8) << A[i][j] << "  ";
+        }
+        cout << endl;
+    }
+}
+
+int main(int argc, char *argv[])
+{
+
+    int size = 128;
+
+    cout << "Initalizing the matricies...";
+
+    random_device rd;
+    mt19937 gen(rd());
+    uniform_real_distribution<> dis(0, 1);
+
+    double *dataA = new double[size*size];
+    double *dataB = new double[size*size];
+    double *dataC = new double[size*size];
+
+    double **A = new double*[size];
+    double **B = new double*[size];
+    double **C = new double*[size];
+
+    for (int i = 0; i < size; i++)	{
+        A[i] = &dataA[size * i];
+        B[i] = &dataB[size * i];
+        C[i] = &dataC[size * i];
+        for (int j = 0; j < size; j++) {
+                A[i][j] = dis(gen);
+                B[i][j] = dis(gen);
+                C[i][j] = 0;
+        }
+    }
+
+    cout << "Done." << endl;
+
+    cout << "Beginning multiply...";
+
+    m5_dumpreset_stats(0, 0);
+    multiply(A, B, C, size);
+    m5_dumpreset_stats(0, 0);
+
+    cout << "Done." << endl;
+
+    delete[] A;
+    delete[] B;
+    delete[] C;
+    delete[] dataA;
+    delete[] dataB;
+    delete[] dataC;
+}