rabbitmq beam.smp cpu利用率过高

问题：
rabbitmq 的吞吐有多大？
目前开启了rabbitmq的持久化/Ack 机制/用的是direct路由器。
上次测试，4.5w/min 的时候，beam.smp cpu利用率过100%
下午做了个测试，不阻塞下游，只测试rabbitmq 的消费速度：

package main
import (
        "fmt"
        "github.com/streadway/amqp"
        "sync"
)
const (
        queueName               = "queue.newpushservice.deal.push"
        exchange                = "exchange.newpushservice.deal.push"
        mqurl                   = "amqp://admin:[email protected]:5672/"
        ReceiverNum         int = 5
        ChannelBufferLength     = 1000
)

func main() {
        fmt.Println("start")
        forever := make(chan bool)
        go WorkMessage(ReceiveMessage())
        <-forever
}

func getChannel() (*amqp.Connection, *amqp.Channel, error) {
        conn, err := amqp.Dial(mqurl)
        if err != nil {
                return nil, nil, err
        }
        channel, err := conn.Channel()
        if err != nil {
                return nil, nil, err
        }
        err = channel.Qos(1, 0, false)
        if err != nil {
                return nil, nil, err
        }
        return conn, channel, nil
}

func ReceiveMessage() <-chan amqp.Delivery {
        out := make(chan amqp.Delivery, ChannelBufferLength)
        var wg sync.WaitGroup

        receiver := func() {
                defer wg.Done()
        RECONNECT:
                for {
                        _, channel, err := getChannel()
                        if err != nil {
                             panic(err)
                        }
                        msgs, err := channel.Consume(
                             queueName, // queue
                             "",        // consumer
                             false,     // auto-ack
                             false,     // exclusive
                             false,     // no-local
                             false,     // no-wait
                             nil,       // args
                        )
                        if err != nil {
                             panic(err)
                        }
                        //检测是否改通道有问题
                        closeChan := make(chan bool, 1)
                        go func() {
                             cc := make(chan *amqp.Error)
                             e := <-channel.NotifyClose(cc)
                             fmt.Printf("[RABBITMQ_CLIENT], channel close error:%s", e.Error())
                             closeChan <- true
                        }()

                        for {
                             select {
                             case msg, ok := <-msgs:
                                if !ok {
                                        fmt.Println("channel be closed,need reconnect")
                                        continue RECONNECT
                                }
                                out <- msg

                             case <-closeChan:
                                continue RECONNECT
                             }
                        }
                }
        }
        wg.Add(ReceiverNum)
        for i := 0; i < ReceiverNum; i++ {
                go receiver()
        }
        go func() {
                wg.Wait()
                fmt.Println("all receiver is done, closing channel")
                close(out)
        }()
        return out
}
func WorkMessage(in <-chan amqp.Delivery) {
        var wg sync.WaitGroup

        worker := func(msg amqp.Delivery) {
                defer wg.Done()
                fmt.Printf("receive message : %s", string(msg.Body))
                msg.Ack(false)
        }
        wg.Add(1)
        go func() {
                defer wg.Done()
                for message := range in {
                        wg.Add(1)
                        go worker(message)
                }
        }()
        go func() {
                wg.Wait()
                fmt.Println("all worker is done, closing channel")
        }()
}

消费速度在 30-40w/min，cpu 占有率在 180-200% ， 内存占用在8%-10%。
4核64位系统,内存16G，cpu 型号（ 8 Intel(R) Xeon(R) Platinum 8163 CPU @ 2.50GHz）
这里有个测试，留着遇到问题可以看
https://segmentfault.com/a/1190000016351345