生產者消費者問題總結

生產者-消費者算是併發編程中常見的問題。依靠緩衝區咱們能夠實現生產者與消費者之間的解耦。生產者只管往緩衝區裏面放東西,消費者只管往緩衝區裏面拿東西。這樣咱們避免生產者想要交付數據給消費者,但消費者此時還沒法接受數據這樣的狀況發生。java

wait notify

這個問題其實就是線程間的通信,因此要注意的是不能同時讀寫。生產者在緩衝區滿的時候不生產,等待;消費者在緩衝區爲空的時候不消費,等待。比較經典的作法是waitnotify編程

生產者線程執行15次set操做安全

public class Producer implements Runnable{
    private Channel channel;

    public Producer(Channel channel) {
        this.channel = channel;
    }

    @Override
    public void run() {
        for(int i=0;i<15;i++){
            channel.set(Thread.currentThread().getName()+" "+i);
        }
    }
}

消費者線程執行10次get操做數據結構

public class Consumer implements Runnable {
    private Channel channel;

    public Consumer(Channel channel) {
        this.channel = channel;
    }

    @Override
    public void run() {
        for(int i=0;i<10;i++){
            System.out.println("Consumer "+Thread.currentThread().getName()+" get "+channel.get());
        }
    }
}

如今定義Channel類,並建立兩個生產者線程和三個消費者線程併發

public class Channel {
    private List<String> buffer=new ArrayList<>();
    private final int MAX_SIZE=10;

    public synchronized String get(){
        while (buffer.size()==0){//不要用if,醒來了也要再次判斷
            try {
                wait();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
        String str=buffer.remove(0);
        notifyAll();
        return str;
    }
    public synchronized void set(String str){
        while (buffer.size()==MAX_SIZE){
            try {
                wait();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
        buffer.add(str);
        notifyAll();
    }

    public static void main(String[] args) {
        Channel channel=new Channel();
        Producer producer=new Producer(channel);
        Consumer consumer=new Consumer(channel);
        for(int i=0;i<2;i++){
            new Thread(producer).start();
        }
        for (int i=0;i<3;i++){
            new Thread(consumer).start();
        }
    }
}

使用notifyAll而不是notify的緣由是,notify有可能出現屢次喚醒同類的狀況,形成「假死」。咱們可使用Condition來實現更精確的喚醒。ide

Condition

將上面代碼中的Channel類修改一下便可this

public class Channel {
    private List<String> buffer=new ArrayList<>();
    private final int MAX_SIZE=10;
    private Lock lock=new ReentrantLock();
    private Condition producer=lock.newCondition();
    private Condition consumer=lock.newCondition();
    
    public String get(){
        String str=null;
        try {
            lock.lock();
            while (buffer.size()==0){
                consumer.await();
            }
            str=buffer.remove(0);
            producer.signalAll();
        }catch (InterruptedException e) {
            e.printStackTrace();
        }finally {
            lock.unlock();
        }
        return str;
    }
    public void set(String str){
        try {
            lock.lock();
            while (buffer.size()==MAX_SIZE){
                producer.await(); 
            }
            buffer.add(str);
            consumer.signalAll();
        }catch (InterruptedException e) {
            e.printStackTrace();
        }finally {
            lock.unlock();
        }
    }
}

雙緩衝與Exchanger

當同步的花銷很是大時,咱們能夠採用雙緩衝區的辦法。雙緩衝的一個好處就在於:由於生產者和消費者各自擁有一個緩衝區,因此他們不會同時對同一個緩衝區進行操做,那麼咱們就不須要爲讀寫操做加鎖,用空間換了時間。在Java中能夠經過Exchanger來交換兩個線程之間的數據結構。線程

public class Producer implements Runnable{
    private List<String> buffer;
    private Exchanger<List<String>> exchanger;
    public Producer(List<String> buffer, Exchanger<List<String>> exchanger){
        this.buffer=buffer;
        this.exchanger=exchanger;
    }
    @Override
    public void run() {
        for(int i=0;i<10;i++){
            for (int j=0;j<10;j++)
            buffer.add("Thrad "+Thread.currentThread().getName()+" : "+i+" "+j);
            try {
                buffer=exchanger.exchange(buffer);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }

    }
}

public class Consumer implements Runnable {
    private Exchanger<List<String>> exchanger;
    private List<String> buffer;

    public Consumer(List<String> buffer,Exchanger<List<String>> exchanger) {
        this.exchanger = exchanger;
        this.buffer = buffer;
    }

    @Override
    public void run() {
        for(int i=0;i<10;i++){
            try {
             buffer=exchanger.exchange(buffer);
            } catch (InterruptedException e) {
             e.printStackTrace();
            }
            for(int j=0;j<10;j++){
                String message=buffer.get(0);
                System.out.println(message);
                buffer.remove(0);
            }
        }
    }
}

public class Main {
    public static void main(String[] args) {
        List<String> buffer1=new ArrayList<>();
        List<String> buffer2=new ArrayList<>();
        Exchanger<List<String>> exchanger=new Exchanger<>();
        Producer producer=new Producer(buffer1,exchanger);
        Consumer consumer=new Consumer(buffer2,exchanger);
        Thread t1=new Thread(producer);
        Thread t2=new Thread(consumer);
        t1.start();
        t2.start();
    }
}

BlockingQueue

咱們可使用更爲方便安全的阻塞式集合來實現生產消費者模型。code

這類集合具備的特色是:當集合已滿或者是爲空的時候,被調用的方法不會當即執行,該方法將被阻塞,直到能夠成功執行爲止。rem

public class Channel {
    private BlockingQueue<String> blockingQueue=new ArrayBlockingQueue<>(10);
    public String get(){
        String str=null;
        try {
            str=blockingQueue.take();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        return str;
    }
    public void set(String str){
        try {
            blockingQueue.put(str);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}

此次的Channel類是否是比以前的簡潔了許多,有了BlockingQueue咱們就不用再去寫wait和notify了。

相關文章
相關標籤/搜索