removed prefetch option from makefile(temporary)

Added tranpose to improve cache performance
run matrix for 800X800 matrix

diff --git a/Robust/src/Benchmarks/Prefetch/MatrixMultiply/MatrixMultiply.java b/Robust/src/Benchmarks/Prefetch/MatrixMultiply/MatrixMultiply.java
index 8c7182534ce7de3450d6b1cb2958959d9d3832f2..8ab2fbb7feb61d004a8768557c0fee02ab35897a 100644 (file)
--- a/Robust/src/Benchmarks/Prefetch/MatrixMultiply/MatrixMultiply.java
+++ b/Robust/src/Benchmarks/Prefetch/MatrixMultiply/MatrixMultiply.java
@@ -16,20 +16,22 @@ public class MatrixMultiply extends Thread{
                atomic {
                    //compute the results
                    localresults=new int[1+x1-x0][1+y1-y0];
-                   
+               
+                   //Transpose b for cache performance
                    for(int i = x0; i<= x1; i++){
                        int a[]=mmul.a[i];
-                       int b[][]=mmul.b;
                        int M=mmul.M;
                        for (int j = y0; j <= y1; j++) {
                            int innerProduct=0;
+                           int b[] = mmul.btranspose[j];
                            for(int k = 0; k < M; k++) {
-                               innerProduct += a[k] * b[k][j];
+                               innerProduct += a[k] *b[k];
                            }
                            localresults[i-x0][j-y0]=innerProduct;
                        }
                    }
                }
+
                atomic {
                    //write the results
                    for(int i=x0;i<=x1;i++) {
@@ -42,7 +44,7 @@ public class MatrixMultiply extends Thread{
        }
 
        public static void main(String[] args) {
-               int mid1 = (172<<24)|(16<<16)|(9<<8)|129;
+               int mid1 = (128<<24)|(195<<16)|(175<<8)|70;
                int mid2 = (128<<24)|(195<<16)|(175<<8)|69;
                int mid3 = (128<<24)|(195<<16)|(175<<8)|71;
                int NUM_THREADS = 2;
@@ -52,7 +54,7 @@ public class MatrixMultiply extends Thread{
                MMul matrix;
 
                atomic {
-                       matrix = global new MMul(70, 70, 70);
+                       matrix = global new MMul(800, 800, 800);
                        matrix.setValues();
                }
 
@@ -62,8 +64,8 @@ public class MatrixMultiply extends Thread{
 
                // Currently it is a 70 X 70 matrix divided into 4 blocks 
                atomic {
-                       mm[0] = global new MatrixMultiply(matrix,0,0,69,35);
-                       mm[1] = global new MatrixMultiply(matrix,0,36,69,69);
+                       mm[0] = global new MatrixMultiply(matrix,0,0,799,300);
+                       mm[1] = global new MatrixMultiply(matrix,0,301,799,799);
                }
 
                atomic {
@@ -84,32 +86,10 @@ public class MatrixMultiply extends Thread{
                System.printInt(r);
                System.printString("\n");
 
-               //Print Matrices to be multiplied
-               /*
-               System.printString("a =\n");
-               for (i = 0; i < p; i++) {
-                       for (j = 0; j < q; j++) {
-                               atomic {
-                                       val = matrix.a[i][j];
-                               }
-                               System.printString(" " + val);
-                       }
-                       System.printString("\n");
-               }
-               System.printString("\n");
-
-               System.printString("b =\n");
-               for (i = 0; i < q; i++) {
-                       for (j = 0; j < r; j++) {
-                               atomic {
-                                       val = matrix.b[i][j];
-                               }
-                               System.printString(" " + val);
-                       }
-                       System.printString("\n");
+               //transpose matrix b
+               atomic {
+                       matrix.transpose();
                }
-               System.printString("\n");
-               */
 
                // start a thread to compute each c[l,n]
                for (i = 0; i < NUM_THREADS; i++) {
@@ -124,26 +104,22 @@ public class MatrixMultiply extends Thread{
                        atomic {
                                tmp = mm[i];
                        }
+                       System.printString("Joining " + i + " ... ");
                        tmp.join();
+                       System.printString("Joined " + i + "\n");
                }
 
                // print out the result of the matrix multiply
-               System.printString("Starting\n");
                System.printString("Matrix Product c =\n");
-               StringBuffer sb = new StringBuffer("");
                int val;
                atomic {
                        for (i = 0; i < p; i++) {
                                int c[]=matrix.c[i];
                                for (j = 0; j < r; j++) {
                                        val = c[j];
-                                       sb.append(" ");
-                                       sb.append((new Integer(val)).toString());
                                }
-                               sb.append("\n");
                        }
                }
-               System.printString(sb.toString());
                System.printString("Finished\n");
        }
 }
@@ -154,6 +130,7 @@ public class MMul{
        public int[][] a;
        public int[][] b;
        public int[][] c;
+       public int[][] btranspose;
 
        public MMul(int L, int M, int N) {
                this.L = L;
@@ -162,6 +139,7 @@ public class MMul{
                a = global new int[L][M];  
                b = global new int[M][N]; 
                c = global new int[L][N]; 
+               btranspose = global new int[N][M];
        }
 
        public void setValues() {
@@ -184,6 +162,20 @@ public class MMul{
                                c[i][j] = 0;
                        }
                }
+               for(i = 0; i < N; i++) {
+                       for(j = 0; j < M; j++) {
+                               btranspose[i][j] = 0;
+                       }
+               }
        }
-}
 
+       public void transpose() {
+               int row;
+               int col;
+               for(row = 0; row < M; row++) {
+                       for(col = 0; col < N; col++) {
+                               btranspose[col][row] = b[row][col];
+                       }
+               }
+       }
+}

diff --git a/Robust/src/Benchmarks/Prefetch/MatrixMultiply/makefile b/Robust/src/Benchmarks/Prefetch/MatrixMultiply/makefile
index c19f69a47b732834bf5ca614355c1726b77cb3cc..3b3016c3f19cdcdbbf8ac97c381fea4c27e29493 100644 (file)
--- a/Robust/src/Benchmarks/Prefetch/MatrixMultiply/makefile
+++ b/Robust/src/Benchmarks/Prefetch/MatrixMultiply/makefile
@@ -1,5 +1,5 @@
 MAINCLASS=MatrixMultiply
 SRC=${MAINCLASS}.java
-FLAGS=-dsm -prefetch -nooptimize -debug -mainclass ${MAINCLASS} -o ${MAINCLASS}
+FLAGS=-dsm -nooptimize -debug -mainclass ${MAINCLASS} -o ${MAINCLASS}
 default:
        ../../../buildscript ${FLAGS} ${SRC}