Nodejs + phantomjs vs. pure phantomjs - page load time

I was poking around recenlty exploring nodejs and phantomjs and wrote a small code to measure the page load time. I have found that page load times differ between the phantomjs code wrapped in nodejs comparing to pure phantomjs code. Below is the code: phantomjs and nodejs for comparison:

Nodejs:

var http = require('http'),
phantom = require('phantom');
url = require("url");

http.createServer(function (request, response) {
  var start = Date.now();

  request.on('end', function () {
    phantom.create(function(ph) {
      ph.createPage(function(page) {
        var _get = url.parse(request.url, true).query;

        page.open(_get[url], function(status) {
          if (status == 'success') {
            var time = Date.now() - start;
            console.log(time);
          }
        });
      });
    });
  });
}).listen(80,'');

Phantomjs:

var page = require('webpage').create();
var system = require('system');

var address = system.args[1];
var time = 0;
var start = Date.now();

page.open(address, function (status) {
  time = Date.now() - start;
  console.log(time + '');
});

The time is usually 4 times longer when testing a site via phantomjs. Any ideas?

PhantomJS doesn't support outputing data, it only open a webpages. phantomjs-node connect them in the following way - it create an instance of ExpressJS and transmits data through its socket.io plus using two more libraries. And data goes:

WebPage > Phantom.js > dnode + node-browserify > Express.js > Socket.io > Node.js

I have the same problem with the speed and performance now reading about node-webkit, thats native WebKit without crutches thats support Node.js modules.