Go典型并发任务

2020-05-23 本文已影响0人 gaobinzhan

仅运行一次

最容易联想到的单例模式：

type Singleton struct {
}

var singleInstance *Singleton
var once sync.Once

func GetSingletonObj() *Singleton {
    once.Do(func() {
        fmt.Println("Create Obj")
        singleInstance = new(Singleton)
    })
    return singleInstance
}

func TestGetSingletonObj(t *testing.T) {
    var wg sync.WaitGroup
    for i := 0; i < 10; i++ {
        wg.Add(1)
        go func() {
            obj := GetSingletonObj()
            fmt.Printf("%x\n", unsafe.Pointer(obj))
            wg.Done()
        }()
    }
    wg.Wait()
    /** 运行结果：
    === RUN   TestGetSingletonObj
    Create Obj
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    1269f78
    --- PASS: TestGetSingletonObj (0.00s)
    */
}

仅需任意任务完成

任务堆里面，只需任务一个完成就返回。

func runTask(id int) string {
    time.Sleep(10 * time.Millisecond)
    return fmt.Sprintf("the result is from %d", id)
}

func FirstResponse() string {
    numOfRunner := 10
    ch := make(chan string) // 非缓冲channel
    for i := 0; i < numOfRunner; i++ {
        go func(i int) {
            ret := runTask(i)
            ch <- ret
        }(i)
    }
    return <-ch
}

func TestFirstResponse(t *testing.T) {
    t.Log(FirstResponse())
    /** 第一次运行结果：
    === RUN   TestFirstResponse
        TestFirstResponse: first_response_test.go:27: the result is from 0
    --- PASS: TestFirstResponse (0.01s)
    */
    /** 第二次运行结果：
    === RUN   TestFirstResponse
        TestFirstResponse: first_response_test.go:27: the result is from 3
    --- PASS: TestFirstResponse (0.01s)
    */
}

因为协程的调度机制，所以返回结果不一样。

但这样是存在很大的问题，修改TestFirstResponse：

func TestFirstResponse(t *testing.T) {
    t.Log("Before:", runtime.NumGoroutine()) // 获取协程数量
    t.Log(FirstResponse())
    time.Sleep(time.Second * 1)
    t.Log("After:", runtime.NumGoroutine()) // 获取协程数量
    /** 运行结果：
    === RUN   TestFirstResponse
        TestFirstResponse: first_response_test.go:28: Before: 2
        TestFirstResponse: first_response_test.go:29: the result is from 6
        TestFirstResponse: first_response_test.go:30: After: 11
    --- PASS: TestFirstResponse (0.01s)
    */
}

因为使用的是非缓冲channel，FirstResponse方法只取走了一次，往channel放入数据的时候，没有被取走，会造成阻塞。

修改非缓冲channel 为缓冲channel就行，否则会造成资源耗尽。

所有任务完成

之前都是用sync.waitGroup实现，这次利用csp机制实现：

func runTask(id int) string {
    time.Sleep(10 * time.Millisecond)
    return fmt.Sprintf("the result is from %d", id)
}

func AllResponse() string {
    numOfRunner := 10
    ch := make(chan string) // 非缓冲channel
    for i := 0; i < numOfRunner; i++ {
        go func(i int) {
            ret := runTask(i)
            ch <- ret
        }(i)
    }

    finalRet := ""
    for i := 0; i < numOfRunner; i++ {
        finalRet += <-ch + "\n"
    }

    return finalRet
}

func TestFirstResponse(t *testing.T) {
    t.Log(AllResponse())
    /** 运行结果：
    === RUN   TestFirstResponse
        TestFirstResponse: all_done_test.go:33: the result is from 9
            the result is from 0
            the result is from 2
            the result is from 7
            the result is from 4
            the result is from 6
            the result is from 1
            the result is from 5
            the result is from 8
            the result is from 3

    --- PASS: TestFirstResponse (0.01s)
    */
}

对象池

使用 buffered channel 实现对象池

image

type ReusableObj struct {
}

type ObjPool struct {
    bufChan chan *ReusableObj // 用于缓冲可重用对象
}

func NewObjPool(numOfObj int) *ObjPool {
    objPool := ObjPool{}
    objPool.bufChan = make(chan *ReusableObj, numOfObj)
    // 提前建立好连接
    for i := 0; i < numOfObj; i++ {
        objPool.bufChan <- &ReusableObj{}
    }
    return &objPool
}

// 获取连接
func (p *ObjPool) GetObj(timeout time.Duration) (*ReusableObj, error) {
    select {
    case ret := <-p.bufChan:
        return ret, nil
    case <-time.After(timeout): // 超时控制
        return nil, errors.New("time out")
    }
}

// 放入连接
func (p *ObjPool) ReleaseObj(obj *ReusableObj) error {
    select {
    case p.bufChan <- obj:
        return nil
    default:
        return errors.New("overflow")
    }
}

func TestObjPool(t *testing.T) {
    pool := NewObjPool(10) // 创建对象池

    for i := 0; i < 11; i++ {
        // 从对象池中获取
        if v, err := pool.GetObj(time.Second * 1); err != nil {
            t.Error(err)
        } else {
            fmt.Println(v)
            // 放入对象池
            if err := pool.ReleaseObj(v); err != nil {
                t.Error(err)
            }
        }
    }

    fmt.Println("Done")
}

sync.Pool对象缓存

image

sync.Pool 对象获取：

尝试从私有对象获取
私有对象不存在，尝试从当前 Processor 的共享池获取
如果当前 Processor 共享池也是空的，那么就尝试去其他 Processor 的共享池获取
如果所有⼦池都是空的，最后就⽤⽤户指定的 New 函数，产⽣⼀个新的对象返回

sync.Pool 对象放回：

如果私有对象不存在则保存为私有对象
如果私有对象存在，放⼊当前 Processor ⼦池的共享池中

sync.Pool 对象生命周期：

GC 会清除 sync.Pool 缓存的对象
对象的缓存有效期为下⼀次 GC 之前

func TestSyncPool(t *testing.T) {
    pool := &sync.Pool{
        New: func() interface{} {
            fmt.Println("Create a new object.")
            return 100
        },
    }

    v := pool.Get().(int) // 从池中获取并断言类型
    fmt.Println(v)        // 100

    pool.Put(3)
    v1, _ := pool.Get().(int)
    fmt.Println(v1) // 3

    //在放进去个 2
    pool.Put(2)
    //为了验证生命周期 这里GC一下
    runtime.GC()
    v3, _ := pool.Get().(int)
    fmt.Println(v3) // 100 而不是 2
    /** 运行结果：
    === RUN   TestSyncPool
    Create a new object.
    100
    3
    Create a new object.
    100
    --- PASS: TestSyncPool (0.00s)
    */
}

func TestSyncPoolMultiGoroutine(t *testing.T) {
    pool := sync.Pool{
        New: func() interface{} {
            fmt.Println("Create a new object.")
            return 10
        },
    }

    pool.Put(100)
    pool.Put(100)
    pool.Put(100)

    var wg sync.WaitGroup

    for i := 0; i < 10; i++ {
        wg.Add(1)
        go func() {
            t.Log(pool.Get())
            wg.Done()
        }()
    }
    wg.Wait()
    /** 运行结果：
    === RUN   TestSyncPoolMultiGoroutine
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 100
    Create a new object.
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
    Create a new object.
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 100
    Create a new object.
    Create a new object.
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
    Create a new object.
    Create a new object.
    Create a new object.
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 100
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
        TestSyncPoolMultiGoroutine: sync_pool_test.go:59: 10
    --- PASS: TestSyncPoolMultiGoroutine (0.00s)
    */
}

sync.Pool 总结：

适合于通过复用，降低复杂对象的创建和GC代价
协程安全，会有锁的开销
生命周期受GC影响，不适合于做连接池等，需自己管理生命周期的资源的池化

Go典型并发任务

仅运行一次

仅需任意任务完成

所有任务完成

对象池

sync.Pool对象缓存

猜你喜欢

热点阅读