Web Browser Grand Prix: Chrome 20, Opera 12, Firefox 13

[alcalde]

The amazing Tom's Hardware Grand Prix lost prestige in my eyes when it decided to not count memory use in the grading. Those who are stating "Memory is there to be used" are the ones whose favorite browser uses so much of it (I won't name names, but it's a semi-official meme on their forum). Memory is there to be used... but not solely by one program, unnecessarily. On modern operating systems, all extra memory is used for caching of previously loaded programs, disk buffers, etc. and has a noticeable benefit on performance. A browser that eats up, and refuses to release, all of this memory degrades the performance of the system. Additionally, few systems are only running one program at a time. Using gobs of memory is going to force other programs to swap memory out to disk, also negatively impacting performance. No one would give Excel or uTorrent a free pass if it grabbed 1.5GB of system memory; why do this for a browser?

I also have to take issue with the claim that browsers are adjusting their memory use based on installed memory; this is another dubious claim put forth by fans of a certain browser that starts with O. Turning off memory caching in that browser still results in the usage of large amounts of memory that is never returned to the OS. It's also probably swapping a lot of memory out to disk.

 

[mayankleoboy1]

@AdamOvera :
i think the IE maze solver test shoulkd not be included in the GP. it hits an outlier in FF, making it look bad. the fact that FF does so good on other CSS tests makes this test a bit suspect.
I think this is the only test where FF is very weak.

 

[adamovera]

[citation][nom]UNKNOWN9122[/nom]If this test were to be done over, they should use an SSD and 8GB 1600MHz RAM. Plus a nVidia GPU for comparison[/citation]
The next test system upgrade will have on all three (Ivy Bridge i5, SSD, DDR3-1600, Nvidia) - probably by fall

 

[adamovera]

[citation][nom]alidan[/nom]hay toms,i use chrome allot, i have probably 100-150 tabs spread out over 4 windows, i never really have more than that, and usualy have less. now i just did a fresh reset of the browser, total system is currently useing 4gb of ram.if i close all chrome, it will use 1.3gb of ramif i leave chrome open, just leave it open, it will bloat to 7-8gb of ram in a few hours. is there any way to benchmark this?[/citation]
Absolutely, I'll look into it for the next one.

 

[adamovera]

[citation][nom]A Bad Day[/nom]What non-MS website uses Silverlight. I uninstalled it like last year and there was never a prompt for a SL installation.[/citation]
Netflix

 

[adamovera]

@alcalde: The browsers do scale total memory usage to the amount of available system memory. While Chrome and IE9 can use up to 1.5GB for 40 tabs on a modern Windows 7 system, they use more like 500MB to display those same 40 tabs on a ten year-old XP machine. So total usage is relative. We now subtract the single tab total from the -39 tab total to see how efficiently the browser manages memory. The lower the difference, the more memory the browser returns to the OS after reducing workload (closing tabs). And I'm really confused by the implication, because w/o counting total usage under load, Opera comes in dead last for memory efficiency...

 

[adamovera]

[citation][nom]mayankleoboy1[/nom]@AdamOvera : i think the IE maze solver test shoulkd not be included in the GP. it hits an outlier in FF, making it look bad. the fact that FF does so good on other CSS tests makes this test a bit suspect. I think this is the only test where FF is very weak.[/citation]
[citation][nom]phate[/nom]Why is Maze solver always such an outlier result for Firefox? And why is it weighted so much. Firefox dominates in the CSS stress test slide (fastest by far). Yet it's result in maze solver puts it in the weak category (I'm assuming) of your results tally. very odd[/citation]
I know it's an FF bug, but I never actually thought it would go unfixed for this long. The problem with removing Maze Solver is that we have no other CSS3 performance test to dispute the assertion that Firefox is weak in CSS3 performance. If anyone knows of another CSS3 performance test, or maybe even a demo Web page created with CSS3 (to which I might be able to apply CSS Stress Testing & Performance Profiling), I'm definitely willing to make a swap.

 

[deltree86]

You guys should try running a .bat with around 400 url's and see which browser does not crash on a mid end system. You will soon see that chrome will crash on xp and somehow firefox will not load all the pages but will not crash your system neither it will crash itself.

 

[Chetou]

[citation][nom]alidan[/nom] firefox 32bit is ok up to 1.25 gb, than it gets choppy, once it hits 1.5gb it becomes borderline unuseableonce it hits 1.7gb, there is no comeing back, need to reset browser.waterfox has no problems like this, i can take it over 3.5gb easily and get very minor preformance problems, keep in mine im only on waterfox 10.[/citation]

Most of my machines are 32bit, so I have to put up with Firefox.

 

[Moneyloo]

[citation][nom]lethalsam[/nom]i won't ever use a browser a browser WITHOUT AD BLOCK Plus. (ABP)ABP works wonderful on Firefox, i RARELY see any ad. While I have used ABP on Chrome BUT its doesn't block half the ads.I know its Not Google's fault, its just that ABP developers are putting more effort with Firefox.So for me, Firefox > Chrome.[/citation]

Try Ghostery with Chrome, haven't seen a single Ad since. Plus as someone who isn't big into social networking, I like that it lets me control what gets sent to those big AD guys. If you use Facebook frequently then make sure you click it's crossed out name when it appears on the icon in the top right of chrome and unblock it, otherwise it will remove the Facebook branding from webpages (which I like).

 

[Guest]

I was an avid Firefox user. I've had several issues in the past that required a rebuild of the profile and reloading all my extensions. I had it running pretty good after that but after the last few updates I started to get repeated lock ups again. The thought of having to rebuild my profile again just gives me tired head. So, I've switched over to Opera and it's been performing much better. Yes, I do miss my FF extensions but I'm getting used to the ones in Opera. Ghostery helps keep your browsing private. And NoAds isn't as good as Adblock but it makes a good attempt. And I've totally gotten hooked on the panel.

 

[dansgas1000]

I have chrome 20 beta, firefox, safari 5 and ie 9

My specs: Onboard grahpics + AMD Athlon 64 X2 5400+ + 2GB RAM. 3.5 on the overall windows score.
Oh and W7 home premium

IE takes 1 min to load up and another 2 mins to open its crappy homepage. It then freezes up the entire f**king thing and i have to close it down.

Firefox is ok, it can be slow at times but i could put up with it for a while.

Safari is quite fast has good add ons.

Chrome is the best its fast has great customization features GREAT add-ons get performance in pretty much everything so i choose that as my fave and i'm using it now

Opera- Havent tried but ive heard its ok.

Thanks for reading.

 

[chewy1963]

FWIW, I use FF 13.0.1 for my default browser and use it for most feature rich type browsing in 5 or 10 tabs.

I use Chrome 20.0.1132.47 m when I want speed and/or >10 of tabs.

Firefox better implementation and more 'compatible' if you will with most heavy content sites.

Chrome better raw speed especially with alot of tabs open.

 

[chewy1963]

Oh ya system specs on main home browser PC:

i3-2120 CPU
HD 2000 graphics
6 GB 1333 mem
1 TB 7200rpm SATA 3g HD
30 Mbit cable internet connection

 

[Guest]

This website is garbage - everytime i click on "drill down", and click back, I'm presented with yet another stupid add smack in the middle of the screen that I have to click off.

 

[Guest]

About acid3 test, did you take into account the not 100% result of some browser ?

 

[jryan388]

ABP is unethical...

 

[jeremyhu]

The "Peacekeeper HTML5 Capabilities" test is a bad test for conformance. At least 3 of the tests I know are not actually testing conformance but rather extensions (which are explicitly *not* counted by html5test.com), and because there are so few "points" in that test, it heavily biases the composite results.

I'd like to see the composite results with that bogus test thrown out.

 

[Guest]

Why would you anchor a fast pc like that with a pathetic spinny hard-drive? The OS + software should be on an SSD, you are years behind the times. And turn off swap while you're at it.

HDD is for media only now.

 

[csf60]

Opera is the only browser without any "weaks" in the essential tests. I guess Opera is a jack of all trades, master of none.

 

[Guest]

M$ IE doesn't have hardware acceleration. Its as slow as shit. Its the slowest browser ever.

 

[Guest]

Woo yeah! Opera!

 

[JacFlasche]

I have never read one of these browser comparison and I didn't read this one. In real estate it is location location location, in browsers it is security security privacy -- Where is the comparison of security and privacy and hackablity? Much more important than a few microseconds or the use of a bit more or less memory. I am still waiting. . . .

 

[adamovera]

[citation][nom]pseudo555[/nom]About acid3 test, did you take into account the not 100% result of some browser ?[/citation]
All five browsers pass Acid3 100%, it was removed from conformance testing in . If that were not the case, we would have gone further back to a couple of Hixie's DOM performance tests. Hopefully, Dromaeo DOM just needs updating.
[citation][nom]jeremyhu[/nom]The "Peacekeeper HTML5 Capabilities" test is a bad test for conformance. At least 3 of the tests I know are not actually testing conformance but rather extensions (which are explicitly *not* counted by html5test.com), and because there are so few "points" in that test, it heavily biases the composite results.I'd like to see the composite results with that bogus test thrown out.[/citation]
It will be, in order to even things out. Yesterday I found a nice new CSS3 conformance test (to finally replace the CSS3 Selectors Test). Next time the conformance composite will have one test each for HTML5, CSS3, and JS.
[citation][nom]Puker[/nom]Why would you anchor a fast pc like that with a pathetic spinny hard-drive? The OS + software should be on an SSD, you are years behind the times. And turn off swap while you're at it.HDD is for media only now.[/citation]
Working on it, the next upgrade will be much sooner than this one was, plan is: i5-3570K, 16GB DDR3-1600, Nvidia Graphics (Fermi, maybe Kepler), and SATAIII SSDs.
[citation][nom]JacFlasche[/nom]I have never read one of these browser comparison and I didn't read this one. In real estate it is location location location, in browsers it is security security privacy -- Where is the comparison of security and privacy and hackablity? Much more important than a few microseconds or the use of a bit more or less memory. I am still waiting. . . .[/citation]
Bottom of . This is first time we could get a security benchmark in an article in time because security benchmarks are difficult to keep relevant. Vendors tend to make sure their products pass them pronto. So, unfortunately, how long this security test lasts is anybody's guess.

 

[adamovera]

[citation][nom]adamovera[/nom]All five browsers pass Acid3 100%, it was removed from conformance testing in .[/citation]
Wow, OK... I guess it's not taking links today. I meant to say "in Web Browser Grand Prix 7." with a link to that article, but no dice.