Question

我有以下代码：

(defn compile-report [id]
    (let [a (gen-first-part id)
          b (gen-second-part id)
          c (gen-third-part id)
          d (gen-fourth-part id)]

    (conj a b c d)))

每个＆x-gen-x-part＆＃39;函数是cpu密集型的。据我了解，let表单将在单个线程上串行运行这些计算。如果我有一个核心机器，那么在它自己的线程上运行它们就没有意义，因为它们都是cpu绑定的。但是，我有一台4芯机器。如何利用每个核心并将这些功能分配到自己的核心？感谢。

Answer 1

您可以将CPU密集型功能包装到future s：

中

buttonVoice.setOnClickListener(new View.OnClickListener() 
{
    @Override
    public void onClick(View v)
    {
        // Get activity from either SearchableInfo or ComponentName
        ComponentName searchActivity = mComponentName;

        // Wrap component in intent
        Intent queryIntent = new Intent(Intent.ACTION_SEARCH);
        queryIntent.setComponent(searchActivity);

        // Wrap query intent in pending intent
        PendingIntent pending = PendingIntent.getActivity(getContext(), 0, queryIntent, PendingIntent.FLAG_ONE_SHOT);

        // Create bundle now because if we wrap it in pending intent, it becomes immutable
        Bundle queryExtras = new Bundle();

        // Create voice intent
        Intent voiceIntent = new Intent(RecognizerIntent.ACTION_RECOGNIZER_SPEECH);
        voiceIntent.putExtra(RecognizerIntent.EXTRA_LANGUAGE_MODEL, RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
        voiceIntent.putExtra(RecognizerIntent.EXTRA_PROMPT, "Speak");
        voiceIntent.putExtra(RecognizerIntent.EXTRA_CALLING_PACKAGE, searchActivity
        voiceIntent.addFlags(Intent.FLAG_ACTIVITY_NEW_TASK);

        // Wrap the pending intent & bundle inside the voice intent
        voiceIntent.putExtra(RecognizerIntent.EXTRA_RESULTS_PENDINGINTENT, pending);
        voiceIntent.putExtra(RecognizerIntent.EXTRA_RESULTS_PENDINGINTENT_BUNDLE, queryExtras);

        // Start the voice search
        getContext().startActivity(voiceIntent);
    }
}

它将使它们在不同的线程中运行。 (defn compile-report [id] (let [a (future (gen-first-part id)) b (future (gen-second-part id)) c (future (gen-third-part id)) d (future (gen-fourth-part id))] (conj @a @b @c @d)))表示@a将阻止，直到结果可用。

如果您的函数处理更大的Clojure数据集，您也可以查看pmap。

如何在Clojure中使用所有CPU内核

1 个答案: