can mpi reduce可以用于java中的字符串数组吗？

Question

我正在从输入文件中进行关于字数统计的课程作业。我的所有同学都使用mpi_send和mpi_recv进行此分配。

虽然我仍然想知道mpi_reduce是否真的不能在这个特定的任务中使用，因为我们已经尝试了但是找不到确切的解决方案。此代码使用hashmap和arraylist。以下是代码：

    //split each line into word-by-word and count the occurrences
    for(int i=0; i < inline.size(); i++) 
    {           
        //to read line by line
        String input = inline.get(i);

        //to split the string on whitespace using regular expression "\\s+"
        String[] words = input.split("\\s+");

    for(int j =0;j<words.length;j++) {
            //start counting the occurences
            for (String word : words) 
            {
                 wordMap.put(word,
                   (wordMap.get(word) == null ? 1 : (wordMap.get(word) + 1)));
            }
            localSum[0]+= words[j];
        }

    }
    int [] globalSum = new int [1];

    //using reduce method
    MPI.COMM_WORLD.Reduce(localSum, 0, globalSum, 0, localSum.length, MPI.OBJECT, MPI.PROD, 0);

    //for time taken
    toc=MPI.Wtime();

    //to ensure all process are completed so that it can proceed to next step
    MPI.COMM_WORLD.Barrier();

    //to display the result at rank 0
    if (rank == 0)
    {
        //display the result
        System.out.println("The results :");
        for (String name: wordMap.keySet()){
            String key =name.toString();
            String value = wordMap.get(name).toString();  
            System.out.println("I am process "+ rank +" "+ key + " = " + value);  
        }
        //display the time
        System.out.println("The time takens :"+ (toc-tic)/100);