Last I looked, almost any objective analysis would give the prize to Apple Safari * Color management correctly applied, or notĬare to guess which Windows browser wins that contest? To look at that in detail, you might consider:
Since this is a photography forum, the speed (and quality) of image rendering is probably a big deal. Think progressive JPEG rendering as just one rather simple example. The manner is which the page is rendered can have a very big impact on perceived performance. The result of all the efforts in all parts of the browser ultimately boils down to the time it takes to grab a page and display it, arguably the primary function of a web browser.Īt the end of the day the fastest browser is the one that displays the results first.Įven that isn't no simple. What about the constantly variable network factors? How many web sites are you going to test to provide a meaningful metric. Note that all your bullets have a say in that metric, depending on what the web designers and devs cooked up.Ĭlick to render is useful but it provides far from a complete picture. Nothing beats the click-to-render time in my book, because it affects everyone, everytime, billions of times a day. This is precisely where we start to get lost in meaningless benchmarks. If you want to test browser performance, you need to use legitimate ways to do it. For one thing, you're testing a specific web site, for another, you're doing so many iterations that caching alone (both internal to a browser, as well as caching from your ISP) is going to have a huge impact on results, not to mention how DNS servers respond for repeat requests for the same addresses they're looking up on a page you're rendering, etc. So, it appears that Microsoft is making some progress.īut, your example of how to test browsers is just not valid for far too many reasons. Now, some of the legitimate benchmarks (and I don't consider you example to be legitimate for a variety of reasons), do show that IE 11 can be faster on some tests (as much as 30% or so faster, not 300% faster).
You'll have network latency issues, DNS server issues, etc.
There are just too many variables involved for that kind of test to be valid.
So, is Firefox really several times as fast as Firefox? I can pit Firefox against Firefox (same version, same machine, same network connection, same DNS server, etc.) and see one several times as fast as the other (even making sure to leave it serial) over tons of repeats.
0 kommentar(er)
