I've decided to move away from HostGator due to their performance and instability lately. I had narrowed my new host down to 2 choices: stablehost and mddhosting. Price and support seem comparable, so now I just wanted to test server speed. I figured if I found a page that both sites host, I can compare accurately. So I found out some server names and ran some tests against their cpanel login page.
Indeed that will not be an accurate measurement.
The only way to truly test would be to have 2 identical sites hosted on each server your are testing even then it still would not be accurate as one server not have as direct route to pingdom.
Good to know that cpanel is not optimized. Although still the hostgator server is downloading 2x as much content in half the time. Does lightspeed really have a 4x effect?
I wish I could host the same site on each server but as you said that's unrealistic. Is there anything else I can use as a benchmark? Just pinging the server or downloading a large file won't showcase what a real website with lots of resources is like.
How am I supposed to choose between hosts based on speed, other than claims?
You can't without signing up and putting a real-world site up there.
What makes a site slow is if the database processes its queries slowly, or the PHP of a complex page on a CMS takes too long to put together.
Neither of those has anything to do with Litespeed, and neither of those can be tested by loading anything other than a page from a fully setup website. To compare two hosts, you'd need the same website on each.
You can't compare in this way. Pick the one you feel most comfortable with on other fronts (support, prices, location, etc.), and sign up. As long as your site is faster than it was with Hostgator, you're fine. If you could have shaved 5% off by switching to the other host on your short-list, you'll never know and you probably won't mind.
The main two ways that I measure web server performance:
1. Install a WordPress with everything set to the defaults. Edit the theme so that the very end of the footer shows the page execution time. I want this to be as low and as consistent as possible. This takes network latency, etc out of the equation so you can see exactly how fast your server is pumping out pages.
2. Use Chrome's developer tools to measure the latency of requests, not just one at a time but also to see how the server copes with dozens of items being loaded at once.