On Speed
Tuesday, March 23, 2010 1:14:14 AM
I noticed some pro-Chrome and pro-Opera bitching in the Desktop team blog comments around performance of Opera 10.51 in the Peacekeeper benchmark. I wrote something to use in a comment there, but then I decided to blog it here instead. It's been a while 
Some choice quotes:
It is pretty much all nonsense to focus on the details and comparing them across systems, to be honest... It is great that the top-score belongs to Opera for now (though the Chrome developers are of course not resting before they get on top again!), and to see Opera in the front lines on all systems. It shows that the JavaScript developers (and other Core and Desktop developers) must be doing something right :-)
But how well a browser really serves you is a highly personal experience. It depends a lot on what kind of sites you visit, what hardware you run your browser on, your Windows version, and what extra browser features you can use/add/customize/get distracted by. The Peacekeeper benchmark seems to be a bit sensitive of your underlying system, in some cases Chrome beats Opera, in others it is the other way around. I don't think speed will be a limiting factor with either of these modern browsers to enjoy the current web.

Some choice quotes:
Originally posted by VarunM:
Chrome 5.0.356.2 destroys Opera 10.51 Especially the "Data" section is ridiculous. Chrome was 300% faster than Opera.
Originally posted by Asires:
"Data" is unimportant in rendering. "Rendering" is important. Opera is twice faster than Chrome in rendering.. So, Opera is 200% faster than Chrome.
It is pretty much all nonsense to focus on the details and comparing them across systems, to be honest... It is great that the top-score belongs to Opera for now (though the Chrome developers are of course not resting before they get on top again!), and to see Opera in the front lines on all systems. It shows that the JavaScript developers (and other Core and Desktop developers) must be doing something right :-)
But how well a browser really serves you is a highly personal experience. It depends a lot on what kind of sites you visit, what hardware you run your browser on, your Windows version, and what extra browser features you can use/add/customize/get distracted by. The Peacekeeper benchmark seems to be a bit sensitive of your underlying system, in some cases Chrome beats Opera, in others it is the other way around. I don't think speed will be a limiting factor with either of these modern browsers to enjoy the current web.






Kyle Bakerkyleabaker # Tuesday, March 23, 2010 2:02:56 AM
Daniel HendrycksDanielHendrycks # Tuesday, March 23, 2010 2:09:24 AM
d4rkn1ght # Tuesday, March 23, 2010 5:30:08 AM
Mağruf ÇolakoğluZAHEK # Tuesday, March 23, 2010 7:14:31 AM
netwolf # Tuesday, March 23, 2010 1:12:02 PM
Originally posted by Rijk:
Very well said
Eric J. Goershappyfunland # Tuesday, March 23, 2010 2:01:12 PM
Kai OckendorfOckendorf # Tuesday, March 23, 2010 8:12:02 PM
Charles SchlossChas4 # Wednesday, March 24, 2010 2:44:59 AM
UtkarshUtkarsh1 # Wednesday, March 24, 2010 6:33:04 AM
Browsing speed, yes, does matter, but it's not the primary base for comparison.
Ice ArdorIceArdor # Wednesday, March 24, 2010 7:28:43 AM
walterbugscout # Wednesday, March 24, 2010 8:09:47 AM
globethrottle # Friday, March 26, 2010 2:51:30 PM
Charles SchlossChas4 # Friday, March 26, 2010 3:25:18 PM
RESULTS (means and 95% confidence intervals)
--------------------------------------------
Total: 354.6ms +/- 1.2%
Jimtoyotabedzrock # Saturday, March 27, 2010 5:08:20 AM
Victor Mogvictor # Saturday, April 3, 2010 9:08:31 AM
szarpaj # Tuesday, April 13, 2010 9:19:16 PM
dapxin # Tuesday, April 20, 2010 6:12:37 AM
Rijk # Thursday, April 29, 2010 7:13:13 AM
Witold Barylukmovax # Saturday, June 26, 2010 5:25:23 AM
benchmarking is very very hard, and even when done right, hard to interpret by people other than browser developers.
Speed is important, not because it allows us run faster existing pages (10ms and 20ms isn't really difference), but by having faster browser we can put considerably more load onto the browser, more complex applications, etc. And this will progress, as having faster browser opens new ways of writing apps and maybe even apps which was condiered before too complex to write (look at jquake2 with gwt in the chrome browser!). There will always be need for speed.
Speed is important, but browser's speed is not only about JavaScript. It is also various forms of rendering (DOM, CSS, SVG, Canvas, images), dns, network, pipelineing, caching, asynchronous loading, compression, etc, itp, and lots of other factors.
QuHno # Saturday, July 10, 2010 7:25:55 PM
May be a new definition of "speed" is needed:
How fast can I achieve what I want by using a certain browser?
That would be / is a very personal speed test. I did it and that is the reason, why I use the same browser since several years (OK, not THE same, here were a couple of snapshots and final versions in between
In my Opera I have rearranged most of the UI elements, I even made my own (Mainly Buttons for I hate Menus in every day use: Button - One Click. Menu - Mostly 2 and a mouse movement) so that they fit exactly to my needs.
Functions I "never" touch are out of sight, all others in plain sight or just 1 (one) click away on a toolbar I blend in with a button (not exactly like in the "12" setup, but similar from the idea). Additionaly I added a ton of shortcuts. All that together still isn't possible with all other browsers I tested, or if, then it was not so easy to do like in Opera.
Speeding me up.
That means speed to me.
Witold Barylukmovax # Wednesday, July 14, 2010 2:26:17 AM
Originally posted by QuHno:
Exactly. This is the reason I use and used Opera always, even in times when Chrome was faster. Becuase UI is so fast, because it is customizable, becuase of mouse gestures, becuase of houndreats of usefull quick tricks.
dovelove # Sunday, August 29, 2010 4:48:48 AM