记录几个测试算法性能的工具类,方便以后测试算法.

Stopwatch类:可以得到算法的执行时间(单位:s,小数点后3位)

   1: public class Stopwatch {
   2:     private long start;
   3:     
   4:     public Stopwatch() {
   5:         start = System.currentTimeMillis();
   6:     }
   7:     
   8:     public double elapsedTime(){
   9:         long now = System.currentTimeMillis();
  10:         return (now - start) / 1000.0;
  11:     }
  12: }
具体使用方法:
   1: public static void main(String[] args) {
   2:         Stopwatch timer = new Stopwatch();
   3:          /*调用某个算法*/
   4:         double spend = timer.elapsedTime();
   5:         /*spend就是调用的算法消耗的时间*/
   6:     }

 

例如:测试选择排序和插入排序的性能

   1: public class SortCompare {
   2:     public static double time(String alg,Comparable[] a){
   3:         Stopwatch timer = new Stopwatch();
   4:         
   5:         if("Insert".equals(alg) || "Insertion".equals(alg)) 
   6:             new Insertion().sort(a);
   7:         else if("Select".equals(alg) || "Selection".equals(alg)) 
   8:             new Selection().sort(a);
   9:         /*此处还可以添加其他的排序算法*/
  10:         
  11:         return timer.elapsedTime();
  12:     }
  13:     
  14:     public static double timeRandomInput(String alg,int times,int length){
  15:         double total = 0.0;
  16:         Double[] a = new Double[length];
  17:         for(int i = 0;i<times;i++){
  18:             for(int j = 0;j<length;j++)
  19:                 a[j] = StdRandom.uniform();
  20:             total += time(alg, a);
  21:         }
  22:         return total;
  23:     }
  24:     
  25:     public static void main(String[] args) {
  26:         String alg1 = args[0];
  27:         String alg2 = args[1];
  28:         int times = Integer.parseInt(args[2]);
  29:         int length = Integer.parseInt(args[3]);
  30:         double t1 = timeRandomInput(alg1, times, length);
  31:         double t2 = timeRandomInput(alg2, times, length);
  32:         StdOut.printf("For %d randomDoubles \n   %s is ", length,alg1);
  33:         StdOut.printf("%.1f times faster than %s\n", t2/t1,alg2);
  34:     }
  35: }
测试结果
p_w_picpath
p_w_picpath
p_w_picpath
p_w_picpath
可以看出结果大致在1~3之间….
 
总结:
在测试或者比较算法的性能时，可以以SortCompare为模板,比较不同算法之间的差异.