Chrome vs Firefox 3.6 vs Safari vs Webkit: Chrome Wins Again | Hacker News Reader

Chrome vs Firefox 3.6 vs Safari vs Webkit: Chrome Wins Again

manu-j.com

18 pointssandosh15 comments

Yeah, but it still renders text funny when css effects such as text-shadow are applied to it. The anti-aliasing goes all early Windows Safari on you.

Cool, keep downvoting. But what I'm saying is that performance etc always comes at the cost of something. In this case: font-rendering and anti-aliasing. Firefox renders text-effects beautifully!

Has anyone else had a problem with Chrome where you hard-drive starts seeking (audible noise of it working) and never stops?

I thought turning off the anti-phishing feature would fix it, but I recently got it again. I've seen it happen on both Macs and Windows PCs.

Yes, if you close the right instance, it stops. but there doesn't seem to be any correlation with whatever page was in that instance. I've also had issues with an instance suddenly using 20-50% of the CPU for no reason.

Chrome is now my primary browser, but this one issue has made me consider switching back to Firefox.

I don't know if this is your problem, but Chrome on Linux has some memory leaks that occasionally put my 2GB desktop into swap. Luckily I can use the Chrome task manager to find the offending tab and kill it.

JavaScript relative performance is still a bit meaningless right now, because market share still counts for more. If you were to write a web app that required the speed of Chrome, it would still suck for most visitors.

This performance mainly showcases Google's platforms (e.g. some netbook with Chrome OS?). Google can write bleeding-edge, speed-dependent web apps, since the whole user experience is based on Chrome.

For everyone else, this doesn't help much yet, but it does mean that the future is bright. Historically Apple and Mozilla have responded well to friendly jabs in benchmarks, so they'll surely work hard to make the next one look much better. And that's when the real fun begins. :)

Actually, these numbers are quite significant to anybody doing mobile web development. Android, iPhone, and WebOS all use Webkit-based browsers.

Javascript performance is particularly relevant for Palm, since all of their applications are JS based.

There's also certain audiences (like HN), where IE is a fractional minority. IE is still keeping us back in the overall market, but there are certain emerging niches where these benchmarks really matter.

Important to view these results in context:

http://v8.googlecode.com/svn/data/benchmarks/v3/run.html

That suite is out of date. They're probably testing using v5, the latest version.

http://v8.googlecode.com/svn/data/benchmarks/v5/run.html

It's great that Javascript is fast and all, but what about the HTML rendering engine? What about the network layer? What about the whole package? I remember a while ago that Safari had an SSL bug that failed to cache certain session data and therefore all SSL sites were dog slow, despite the rendering engine being fast.

Is anyone doing any real-world benchmarking with actual websites? But since throwing in the Internet to the mix would throw off the results, perhaps a locally hosted web server to do the actual serving? Create a clone of several popular sites locally, serve them up to the browsers of choice, time the results, ???, profit!

Curiously? Does anyone know why IE8 is left out of comparisons like this? I'd love to know how it ranks for JS performance against the "other" browsers

According to all the benchmarks I've seen, IE8 is about an order of magnitude (10x) slower than any of the Webkit browsers. They're so far behind that you'd have to significantly rescale the graphs to include them.

Here's an example of a recent Sunspider benchmark: http://www.computerworld.com/s/article/9138459/IE8_runs_10_t...

John Resig has a performance rundown that is about one year old that included IE (http://ejohn.org/blog/javascript-performance-rundown/). IE is woefully slow compared to the other browsers in the roundup and is not included in some of the results as it would continually crash during testing. It would be nice to see some more up to date benchmarking with IE, but the link I referenced can give you some insight as to its speed relative to the other browsers.

It usually cannot complete benchmarks that bring out the differences between the good engines.

is it really newsworthy when benchmarks reiterate the status quo?

Reply on news.ycombinator.com