Parallel array summation

C++ code:

#include <stdio.h>#include <stdlib.h>#include <pthread.h>#define NUM_THREADS 4int N;int *X;int gSum[NUM_THREADS];//pthread_mutex_t sum_mutex = PTHREAD_MUTEX_INITIALIZER;// do sumvoid *summation(void *pArg){int tNum = *((int *) pArg);int lSum = 0;int start, end;start = (N / NUM_THREADS) * tNum;printf("start is %d\n", start);end = (N / NUM_THREADS) * (tNum + 1);printf("end is %d\n", end);if(tNum == (NUM_THREADS - 1)) {end = N;}for (int i = start; i < end; i++){lSum += X[i];//printf("%d ", i);}//pthread_mutex_lock(&sum_mutex);gSum[tNum] = lSum;//printf("%d ", gSum[tNum]);//pthread_mutex_unlock(&sum_mutex);free(pArg);}// initialize arrayvoid initializeArray(int N){X = new int[N];int i;for (i = 0; i < N; i++){X[i] = i + 1;//printf("%d ", X[i]);}}int main(int argc, char* argv[]){int j, sum = 0;pthread_t tHandles[NUM_THREADS];N = 5;initializeArray(N);for (j = 0; j < NUM_THREADS; j++){int *threadNum = new(int);*threadNum = j;pthread_create(&tHandles[j], NULL, summation, (void *)threadNum);}for (j = 0; j < NUM_THREADS; j++){pthread_join(tHandles[j], NULL);sum += gSum[j];//printf("%d\n", gSum[j]);}printf("The sum of array elements is %d\n", sum);}

Java code:

package hello.test;import java.util.Random;public class ThreadSum extends Thread {private final int R = 10000;// the random number thresholdprivate int NUM_THREADS = 4; // number of threadsprivate int[] a; // array that to sumprivate int N;// the number of the arrayprivate int start;private int end;private static long[] gSum; // global array to store array chunk sumprivate int ithread; // the ith thread number//private int sum;// constructorpublic ThreadSum(int N) {this.N = N;a = generateRand();gSum = new long[NUM_THREADS];for (int i = 0; i < gSum.length; i++) {gSum[i] = 0;}}// constructorpublic ThreadSum(int[] a, int start, int end, int ithread) {this.a = a;this.start = start;this.end = end;this.ithread = ithread;}// generate the random arrayprivate int[] generateRand() {// TODO Auto-generated method stubRandom r = new Random();int[]a = new int[N];for(int i = 0; i < a.length; i++) {//a[i] = r.nextInt(R);a[i] = i + 1;}return a;}// sum all the elements of array with threadspublic void summation() {int start, end;ThreadSum[] ts = new ThreadSum[NUM_THREADS];for (int i = 0; i < NUM_THREADS; i++) {start = (N / NUM_THREADS) * i;end = (N / NUM_THREADS) * (i + 1);if(i == (NUM_THREADS - 1)) {end = N;}ts[i] = new ThreadSum(a, start, end, i);ts[i].start();}for (int j = 0; j < NUM_THREADS; j++) {if(ts[j] != null) {try {ts[j].join();} catch (InterruptedException e) {// TODO Auto-generated catch blocke.printStackTrace();}}}}// sum the array which index is from start to endprivate void summation(int start, int end) {long lSum = 0;for (int i = start; i < end; i++) {lSum += a[i];}//return lSum;gSum[ithread] = lSum;}// multithread override public void run(){summation(start, end);}// return globla array gSum which stored the sub sum of array chunkspublic long[] getgSum() {return gSum;}// return the number of threadspublic int getNumThreads() {return NUM_THREADS;}// get threaded sumpublic void getThreadSum() {long[] result = getgSum();long sum = 0;for (int i = 0; i < result.length; i++) {sum += result[i];System.out.println(" result[" + i + "]: " + result[i]);}System.out.println("The thread sum is : " + sum);}// get serial sumpublic void getSerialSum() {long sum = 0;for (int i = 0; i< a.length; i++) {sum += a[i];}System.out.println("The serial sum is : " + sum);}// test functionpublic static void main(String[] args) {ThreadSum threadsum = new ThreadSum(100000);long startTime1 = System.nanoTime();threadsum.summation();threadsum.getThreadSum();long stopTime1 = System.nanoTime();System.out.println("threaded sum cost time1 is: " + (stopTime1 - startTime1));long startTime2 = System.nanoTime();threadsum.getSerialSum();long stopTime2 = System.nanoTime();System.out.println("threaded sum cost time1 is: " + (stopTime2 - startTime2));}}

Java code 2:

package hello.test;import java.util.Random;public class ThreadSum {private final int R = 10000;// the random number thresholdprivate int NUM_THREADS = 4; // number of threadsprivate int[] a; // array that to sumprivate int N;// the number of the arrayprivate int start; // start index of array chunkprivate int end; // end index of array chunkprivate static long[] gSum; // global array to store array chunk sumprivate int ithread; // the ith thread number// constructorpublic ThreadSum(int N) {this.N = N;a = generateRand();gSum = new long[NUM_THREADS];for (int i = 0; i < gSum.length; i++) {gSum[i] = 0;}}// constructorpublic ThreadSum(int[] a, int start, int end, int ithread) {this.a = a;this.start = start;this.end = end;this.ithread = ithread;}// generate the random arrayprivate int[] generateRand() {// TODO Auto-generated method stubRandom r = new Random();int[]a = new int[N];for(int i = 0; i < a.length; i++) {a[i] = r.nextInt(R);//a[i] = i + 1;}return a;}// sum all the elements of array with threadspublic void summation() {int start, end;for (int i = 0; i < NUM_THREADS; i++) {start = (N / NUM_THREADS) * i;end = (N / NUM_THREADS) * (i + 1);if(i == (NUM_THREADS - 1)) {end = N;}Thread thread = new Thread(new SumRunnable(a, start, end, i)); thread.start();try {thread.join();} catch (InterruptedException e) {// TODO Auto-generated catch blocke.printStackTrace();}}}// sum array chunk from index start to endprivate void summation(int start, int end) {int lSum = 0;for (int i = start; i < end; i++) {lSum += a[i];}//return lSum;gSum[ithread] = lSum;}// return globla array gSum which stored the sub sum of array chunkspublic long[] getgSum() {return gSum;}// get threaded sumpublic void getThreadSum() {long[] result = getgSum();long sum = 0;for (int i = 0; i < result.length; i++) {sum += result[i];System.out.println(" result[" + i + "]: " + result[i]);}System.out.println("The thread sum is : " + sum);}// get serial sumpublic void getSerialSum() {long sum = 0;for (int i = 0; i< a.length; i++) {sum += a[i];}System.out.println("The serial sum is : " + sum);}// inner class that implement runnable interfaceprivate class SumRunnable implements Runnable {private int[] array; // the array to be sumprivate int start;// start indexprivate int end; // end indexprivate int ithread; // ith thread number// constructorpublic SumRunnable(int[] array, int start, int end, int ithread) {this.array = array;this.start = start;this.end = end;this.ithread = ithread;}// override run()public void run() { ThreadSum threadsum = new ThreadSum(array, start, end, ithread);threadsum.summation(start, end);}}// test functionpublic static void main(String[] args) {ThreadSum threadsum = new ThreadSum(1000000);long startTime1 = System.nanoTime();threadsum.summation();threadsum.getThreadSum();long stopTime1 = System.nanoTime();System.out.println("threaded sum cost time1 is: " + (stopTime1 - startTime1));long startTime2 = System.nanoTime();threadsum.getSerialSum();long stopTime2 = System.nanoTime();System.out.println("threaded sum cost time1 is: " + (stopTime2 - startTime2));}}

Conclusion:  

The serial array summation is much faster than threaded array summation with small array numbers. When the array number is much bigger,

The cost time is get closer and closer.