I would suggest using the "headless" PhantomJS (which is related to Selenium WebDriver) or using the Selenium HTMLUnitDriver. Wrap the load tests in a JUnit class. Fork the tests multi-threaded using either Maven Surefire or Gradle (with a filtering test task using maxParallelForks). Gradle or Maven will create a JUnit report (from a CSS transformation) that shows each test and the time each one took.
If you are using Windows, you can can use my .bat scripts to start a Grid Node and Hub in order to run parallel, using multiple computers to supply the load..
