Java parallel programming to sum long number

以下代码是我写的，思路基于 上一篇，再次证明 for 循环有点耗费时间。一旦使用了for 循环创建线程，我们无法得到大于 1 的加速比, 也就是说，平行化失败！

   class MyTimer {
    private static long begin;

    public static void time_start(){
        begin = System.nanoTime();
    }

    public static double time_end_milli_seconds(){
        return System.nanoTime()-begin/1e6;

    }

    public static double time_end_seconds(){
        double total_time =  (System.nanoTime()-begin)/1e6;
        return ((double)((int)total_time))/1000;

    }

}

class MySumThread extends Thread{
    public long sum = 0;
    public static int cpu_num = Runtime.getRuntime().availableProcessors();

    public MySumThread(long begin_index) {
        this.begin_index = begin_index;
    }

    public long begin_index;

    @Override
    public void run() {

        for (long i = begin_index; i < 1000000001; i+=4) {
            sum+=i;
        }
    }
}

public class MySumTask{

    public static long sum_result = 0;

    public static int cpu_num = Runtime.getRuntime().availableProcessors();

    public static double linear_sum(){

        MyTimer.time_start();

        long sum = 0;
        for (long i = 0; i < 1000000001; i++) {
            sum+=i;
        }

        double end_time = MyTimer.time_end_seconds();
        System.out.println(" linear_sum : "+sum+" time elapse : "+end_time);

        return end_time;

    }

    public static double parallel_sum() throws InterruptedException {
        MyTimer.time_start();

/*
        for (int i = 0; i < cpu_num; i++) {
            MySumThread thread = new MySumThread(i+1);
            thread.start();
            thread.join();
            sum += thread.sum;
        }  //it's too slow ...the speed up is over 0.6
*/

        MySumThread thread1 = new MySumThread(1);
        thread1.start();

        MySumThread thread2 = new MySumThread(2);
        thread2.start();

        MySumThread thread3 = new MySumThread(3);
        thread3.start();

        MySumThread thread4 = new MySumThread(4);
        thread4.start();


        thread1.join();
        thread2.join();
        thread3.join();
        thread4.join();

        sum_result += thread1.sum + thread2.sum + thread3.sum + thread4.sum;

        double end_time = MyTimer.time_end_seconds();
        System.out.println(" parallel_sum : "+sum_result+" time elapse : "+end_time);
        return end_time;
    }

    public static void main(String[] args) throws InterruptedException {

       System.out.println("speed up : "+linear_sum()/parallel_sum());

    }

}