go语言使用goroutines和channel实现一个工作池相当简单。使用goroutines开指定书目线程，通道分别传递任务和任务结果。简单的线程池代码如下：

1 package main 2  3 import "fmt" 4 import "time" 5  6 // Here's the worker, of which we'll run several 7 // concurrent instances. These workers will receive 8 // work on the `jobs` channel and send the corresponding 9 // results on `results`. We'll sleep a second per job to10 // simulate an expensive task.11 func worker(id int, jobs <-chan int, results chan<- int) {12     for j := range jobs {13         fmt.Println("worker", id, "started  job", j)14         time.Sleep(time.Second)15         fmt.Println("worker", id, "finished job", j)16         results <- j * 217     }18 }19 20 func main() {21 22     // In order to use our pool of workers we need to send23     // them work and collect their results. We make 224     // channels for this.25     jobs := make(chan int, 100)26     results := make(chan int, 100)27 28     // This starts up 3 workers, initially blocked29     // because there are no jobs yet.30     for w := 1; w <= 3; w++ {31         go worker(w, jobs, results)32     }33 34     // Here we send 5 `jobs` and then `close` that35     // channel to indicate that's all the work we have.36     for j := 1; j <= 5; j++ {37         jobs <- j38     }39     close(jobs)40 41     // Finally we collect all the results of the work.42     for a := 1; a <= 5; a++ {43         <-results44     }45 }

　　执行上面代码，将得到以下输出结果

1 worker 3 started  job 2 2 worker 1 started  job 1 3 worker 2 started  job 3 4 worker 3 finished job 2 5 worker 3 started  job 4 6 worker 1 finished job 1 7 worker 1 started  job 5 8 worker 2 finished job 3 9 worker 3 finished job 410 worker 1 finished job 5

　　看代码中注释也大概能理解每一步的含义，代码25行和26行分别初始化了2个通道，用于发送任务给子线程和接收子线程计算的任务结果。30-32行代码是启动了3个子线程，用于处理任务，并将任务通道和结果通道传递给了线程函数。36-38行代码是发送任务到jobs通道，工作线程在没有任务时，是阻塞着等待任务，当发现任务通道中有任务时，开始执行任务，当任务执行完毕时，将任务结果发送给结果通道。

　　jobs <-chan int：只能接收数据

　　results chan<- int：只能发送数据