Web Browser Grand Prix 5: Opera 11.50, Firefox 5, And Chrome 12

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
[citation][nom]a-nano-moose[/nom]You should include the time it takes to reload the failed pages in the 40 tab test. It should be the time it takes to load 40 tabs right.[/citation]
No, each broken page was re-loaded individually, if we reload all 40 at once, we'll always end up with broken pages on the browsers which failed in the first place.
 
[citation][nom]some body[/nom]The scoring methodology could be improved. Assigning 1,2,3,4,5 points based on the encog silverlight timings which range between 6.359522 seconds and 6.4928713 is ridiculous. Just showing that much precision is ridiculous. Any timings which are within 1% of another should be considered a tie.[/citation]
That's just the scale of Encog. When we started with that benchmark all the browsers earned scores of 6.49xxxx. A better scale from the test would be nice though. If a newer, better Silverlight benchmark comes out, we'll switch.
 
I have no allegiance to a browser and use IE, Firefox, and Chrome. No browser really seems to stand out with a particular feature set other than I find myself turning to IE when pages aren't functioning correctly in the other browsers.
 
After using FF5 for sometime, task-manager shows FF5 used ~250M. But when I closed FF5, I saw the total memory usage dropped about 500M. Where are the hidden memory usage comes from?
 
I was a die hard FireFox user, since before it was actually FireFox. A few years back saw the advent of Chrome, and it's since become my primary browser. I also like IE9 as well, though change the homepage to google -- I was not surprised to see how long it took to load MSN.com, a godawful 2.5 seconds. I like that there's a lot to choose from at the moment.

Firefox was the greatest browser, bar none, for quite some time. Without Firefox there certainly wouldn't be the WBGP or the browser as we currently know it. It surely isn't the best for anything except the add on department at the moment. Mozilla needs to continue to play their game, not capitulate to the absurd Chrome development cycle naming convention. Most of all Firefox isn't the only reasonable alternative to IE; worse, IE is pretty good now.

I say its a good thing that all of the browsers (save Safari) are basically pretty good. Each has something different to offer, and I do like to switch up after a while.
 
Can we please begin every review like this with a disclaimer that these benchmarks are no good reason to choose a browser? I use Firefox because of the IR L337 extensions, but in any normal usage pattern on any reasonably up-to-date PC (Intel Atom not withstanding), they all feel just about the same in real world use. I occasionally use Chrome or IE at work, and they aren't faster in any meaningful way.
 
[citation][nom]JOSHSKORN[/nom]Now if just Google would release a 64-bit Chrome browser.[/citation]
most web pages are not coded for 64bit, therefore will not load properly and some will even fail to load
 
Sure looks like Firefox is quickly becoming the IE 6 of the decade. Somehow they have lost their way. Google for the most part is king right now. IE 9 is at least very respectable. Opera looks good too,but unfortunately nobody really cares.
 
[citation][nom]adamovera[/nom]No, each broken page was re-loaded individually, if we reload all 40 at once, we'll always end up with broken pages on the browsers which failed in the first place.[/citation]
I should have worded my statement better. That is what I meant. Keep the timer going as you keep reloading the individual broken tabs until they succeed.
 
I think one of the other tests Toms must perform that has been left out.
* How long it takes to load up 15 tabs and unload.. then re-load them back.

This will bring a good real world example. I will do this a lot, (perhaps not 15 pages.. but up and down at least 3 or 4) so it is a very good test.

What I found.(real life). firefox is slow to start.. and I noted chrome is nice and fast.. same as IE 9. I tried IE9 and Chrome. I enjoyed the speedy start-up of both, but after loading up more tabs.. and removing and reloading.. the browser speed was much slower than firefox or opera. so the user experience was not as good, (apart from the initial start-up). so I reverted back to Firefox. with Opera as the backup.

A further note..My take on IE9
IE9 is not that good at handling multi tab page loading, very unstable with some complex web pages and hence very unsatisfactory for the user experience.
I am sure it will improve, but for a single operating system browser.. it is VERY VERY sad to see such a poor product. I expect a single operating system browser to be 10x better as they do not have to suppport multiple operating systems.

So I am not impressed that Tom's gave IE9 a high rank.. Take some points off for lack of operating system support.

my 2c's
 
I'm starting to think that my favorite browser, Safari, is just a piece of crap under Windows 7. I don't get it. I wonder how these browsers stand against each other under Mac OS X. I'm having hard time to believe these results. Under Mac OS X it seems all is just opposite.
 
[citation][nom]adamovera[/nom]We tried emptying the bin in WBGP3 (or 4?), it had no effect whatsoever. But I'll try again tomorrow before I wipe the test HDD.[/citation]

I wouldn't bother. This is an urban legend the Opera faithful have been telling themselves for years to try to deny that the memory bug exists. It doesn't work and never has. The other story you'll hear is the nonsensical "if other programs need the memory, Opera will give it back". There's no function in any OS for programs to ask other programs for memory, of course. When you point that out, the next story will be "Memory is there to be used!" Yes, by the OS, for disk and library caches, not Opera.

Opera's bug tracker is "secret", so despite people submitting bug reports for years now, Opera doesn't acknowledge whether they even consider this a bug or not. They don't respond to any questions about how their memory management (doesn't) work, and of course they don't acknowledge or comment on any of Tom's browser battles. They let volunteers moderate their forums (!), and these volunteers find excuses to close and/or delete any thread in which this subject comes up - often with the help of an accomplice who will intentionally use an expletive in a thread to give them the excuse they need. It's ridiculous.

The memory problems are even worse under Linux since Opera uses much more memory in that version. I've found that Opera apparently loads the full web pages that are in its speed dial, because loading Opera with no tabs open but 12 speed dial entries results in the use of 380MB on start-up! Even without speed dial, my tests were still able to leave a situation with no tabs open and 500MB in use (tab recycle bin emptied and memory cache set to off).

Tom's Hardware is the Oprah Winfrey of websites. 🙂 You are too powerful to be denied. I'd weep for joy if you guys and gals were to demand an interview with Opera folks, pretend they were a faulty Intel CPU and grill'em! 🙂 Demand answers - some people have actually had to create their own blogs to discuss Opera issues because the company's own forums were too hostile. You could get them to admit the issue, force them to explain what the heck their browser (isn't) doing, demand to know why they ignored the problem for about two years, make them declassify their bug tracker, bring freedom to the forums, accuse them of chasing Chrome's benchmarks and Firefox's extensions and ignoring Opera's... operishness, and get things back on track! It's not even just the memory issue; that's one of many problems. Auto-completion of .net, .com, .edu stopped working when they made the URL bar into a search bar like Chrome (even though Opera already had a search box so it was just trying to work like Chrome). Forum posts reporting this got no official response (but attacks from the zealots). Filing bug reports got no acknowledgement because of the "secrecy". Now after several versions, the option is STILL there to do auto-completion in the settings! So is it a bug then? Are they going to (someday) give us an option (after more than a year) to go back to the previous behavior? Or will nothing ever happen because they're happy with things as they are? With Opera, we the users never know.

I've been an Opera user since the days of paying for it, but I'm about to switch browsers because it's just ridiculous. Opera is trying to be Chrome and Firefox rather than Opera and bugs go unfixed and cool new features have stopped appearing and really nice ones (like the Opera Unite in-built server) wither on the vine with no new enhancements or development.

Tom's Hardware, like Obi-Wan Kenobi, you're our only hope! Only you can get answers for we the ignored, abused, censored (former and soon-to-be-former) users of Opera!
 
Congratulations on the fantastic report!
You are adopting the scientific method more and more, which only contributes to the objectivity of the report. You measure many different things and control for a lot of independent variables.
I would suggest that you apply some normalization to the performance measurements as well, in some charts the differences between browsers are within 1-2%, while in other it's several hundred percent. In this case we have very different first or last places. This would make for an even more accurate benchmark. Looking forward to seeing more awesome work from your team :)
 
[citation][nom]lucas1024[/nom]Thanks for including the reliability test! I had almost convinced myself to switch to Chrome from FF5, because of the memory usage, but not anymore - I do like my pages to load every time and I routinely have 30-40 tabs open.[/citation]

30-40?!?!?!? Holy cow, I can't stand more than 5 open at once!
 
LINUX Support How can you say IE is comparable to the browsers that support LINUX. Every user that uses linux cannot use IE. Until they port to IE how can a business standardise on IE if it does not support all platforms. ?
 
Fast is useless if the results are garbage. In my world IE = Incompetently Engineered. A professional web designer I *still* have to spend hours working around IE bugs -- and it doesn't help that most Windows users can't upgrade to IE9, thanks to the marketing department...
 
I find it hilarious that IE9 loads msn.com, it's own default homepage, at almost 2.5 seconds while Chrome, Safari and Safari take less than 1 second. Shame on you IE9!

For me, speed and security are key. That's why I use Google Chrome.
 
I have something like 180 tabs open right now in FF5 and I seriously don't care about memory usage since I have 2 GB of RAM. I'm actually thinking of sandboxing my browser inside a virtual machine, on a lightweight Linux distribution with nothing but the X window system, running in full screen. Allowing it around 768 MB RAM should be enough, with the .vdi file around 500 MB or so, for cache purposes (probably a bit much but who knows how much "video" I'll view in a session 😀).

Haven't used IE in years. Maybe the occasional default (virtual machine) install to look something up (extremely rare situation) or at a newbie friend's house, or school or something. If Microsoft wants to gain market share with IE I think they should seriously consider naming it something else, because IE6 from the XP days gave it a very very bad name.
 
Chrome has become an impressive performer in short order. It still has some irritating user-interface quirks: the ESC key won't stop moving GIFs, the bookmarks list doesn't scroll at the speed configured on a wheel mouse or trackball, and AdBlock is not as good as on Firefox. I find myself using it more and more, however.
 
I am a little caught of guard by the page load errors found on chrome. I switched from Firefox to Chrome well over a year ago and I believe I could count on one hand the times I've had to reload a page.
 
[citation][nom]jacob0808[/nom]I am a little caught of guard by the page load errors found on chrome. I switched from Firefox to Chrome well over a year ago and I believe I could count on one hand the times I've had to reload a page.[/citation]

That's true, I remember looking at the bug report from the Chromium project for reported page load errors due to premature timeouts and the google engineers basically stated that they wouldn't fix the problem in the name of keeping Chrome speedy. The problem is more pronounced if you visit sites based oversees or have a slow internet connection.
 
"There's no argument; Chrome 12 is the Web Browser Grand Prix 5 champion."

I always get the feeling that for many of these articles, the conclusions are in "draft form" before the testing begins. Just as a lark, I assigned 4 points for "winning", 3 for "strong", 2 for "acceptable" and 1 for "weak". The results if the math I did in my head was right were Chrome 53, Firefox 53, IE 50, Opera 51, .

While each user may rank each of the categories differently, combining this observation with the responses, sounds like the race is closer than the quoted sentence above would indicate.
 
If you'd like to see something interesting, then download Chromium and run it against http://html5test.com/ along with Chrome, they get different scores.
 
Status
Not open for further replies.