I wanted to test the performance of concurrent http request in Go against node.js:

package main

import (

var responseCounter = 0
var requestCounter = 0
var count=0
var ch = make(chan int)

func sendRequest(){
    go func() {
        url := "https://www.google.co.in/#q=search_" +strconv.Itoa(requestCounter)
        resp, err := http.Get(url)
        if err != nil {
        defer resp.Body.Close()
        ch <- count

func main() {

    for i := 1; i<100; i++{

    for {
        select {
            case r := <-ch:
                fmt.Printf("\nChannel ",r)
            case <-time.After(50 * time.Millisecond):

When I run this code the CPU usage goes very high (around 90%). Is there something wrong with this code? Have I used the Goroutines correctly?

  • \$\begingroup\$ but you start infinitely many goroutines in your sendRequest, no? \$\endgroup\$ – akonsu Jan 9 '15 at 16:29
  • \$\begingroup\$ I did so because I wanted the requests to be continuous. So when one http request is complete I made another request by calling sendRequest(). Is there a better way to achieve this? \$\endgroup\$ – Sarita Jan 10 '15 at 15:52
  • \$\begingroup\$ try to wait for the goroutine to finish and then start another one. I do not know what is going on... \$\endgroup\$ – akonsu Jan 10 '15 at 16:19
  • \$\begingroup\$ So many data races!! (go {build,install,run,test} -race is your friend). \$\endgroup\$ – Dave C Apr 12 '15 at 17:00

Inside SendRequst you are calling SendRequest again, which will make your stack full of recursive calls.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.