i'm not surprised.

there are a lot of programs that run faster and better in windows, than those that microsoft makes.

Well I don't know how you did your tests - but it just didn't sound right to me so I did some simple real world tests, namely opening 12 tabs simultaneously in Firefox and Chrome in Win7. The tests were done on freshly opened browsers. My Firefox is loaded up with a whole bunch of add-ons (16 with the same add-ons in XP). Same browser versions for Firefox (3.0.11). For hrome I have the Beta (3.0.187.1) in 7 and in XP I have the stable (2.0.176.0).
To put it simply THEY ARE ALL THE SAME.
Note I actually had one less site to load in XP but even so Firefox recorded the slowest load time for the 11 sites in XP A WHOLE THREE SECONDS SLOWER than the fasted load time (which was Firefox in Win 7.
Basically NONE of your results were replicated in my real world situation. Despite claims Chrome was slower than FF in 7 and 2 secs faster in XP.
I guess chrome wins because I have Ad-Block+ in all my FF installs wheras Chrome had to load ads!

Win 7 (12 tabs)

FF..........45secs
Chrome...47secs

XP SP3 (11 tabs)

FF..........48secs
Chrome...47secs

So my truth is that both FF and Chrome are good in Win 7 with Firefox the fastest overall. Chrome is NOT the fastest browser on the planet. The differences overall are trivial.
(BTW - the sites were mainly Australian shopping sites with a couple of US sites. Heady mix of flash, JavaScript, Images, text and Tables. When the last spinner stopped I clicked the stopwatch.)

So you like webkit, but I still use FF, what is the point? Do you think Windows Vista 7 is faster than XP? Do users care? I don't. I know that XP just works and I don't want to experiment. Microsoft whould scrap Windows 7 and release some interface updates for XP.

*sigh*

Ok we get it.. you LOVE CHROME.. just do an opinion piece and state it.. this neverending stream of essentially useless comparisons.. will do nothing to change the gamefield in regards to browsers..

IE is dominant because it works well enough to get what the mass market needs done, and it is already on the machine..

Firefox is #2 because it blocks ads (take adblock+ out of firefox and the marketshare would be on par with the rest of the single digit browsers) it also does pretty much everything else you would want to do with a browser as well as the standard browser.

UNTIL you can do what firefox does, better than firefox.. you wont supplant firefox marketshare..

I could put out a browser tommrrow that renders everything 100% according to standards.. instantly.. but that will not block ads (since this theoretical browser is being financed by the doubleclick network and aol) .. performance wont matter.. because the net gain to the user is nil.. they still see ads, and they have to install and learn new software.. end result? Opera level marketshare.

And google is not likely to EVER allow a real adblock solution in the Google Branded Chrome.. (hopefully the people that take the engine and strip it of google's phone home/ad loving nonsense are more successful)

Call me stupid, but can I ask something. Does speed really matter in the modern age? Yes, an application should be quick to run and perform, but will savings .008ms on a page load really make the difference between whether or not you will use the application? Will you swap because opera 10 build 103781 is .002microseconds faster than your current browser?

I have seen a lot of these topics lately from BN and I am beginning to wonder what suddenly brought them into the utmost importance that we need 8-9 new topics per week about the subject?

A valid question to be sure, but the speed goals of these latest techniques (mostly in chrome/safari) are not meant to speed up the overall render time of a given page..

They are meant to make a web app such as gmail, or a web word processor, anything that runs within a browser entirely.. respond more like it is a regular compiled app..

Historically speaking while you could make a very fancy complicated web app UI that mirrored a traditional app, the response time actually USING the app was hideous because every time you did something.. the UI had to be run through code meant to add simple fairly static objects to webpages (stuff like a counter, or slideshow that loaded once and was left alone for the rest of the page view)

the problem we face with these betanews silly posts touting how fast chrome/safari/etc are.. is that as others have pointed out.. its purely theoretical, and even where there is a compelling margin of increase in the real world.. these tests are for a very specific subset of the actual work we all do with browsers on a daily basis.

You see, that I fully understand. I realize what most browser makers are trying to do and I see where the speed increases are most realized in each browsers tweaking method. But is that mentioned anywhere on these BN posts or is this subset of new "news articles", which state essentially the same thing time after time, strictly focused on how fast a browser can render various html and JS pages? The problem, as I see it, is as you and others have mentioned. There is no real way to validate these results. Just because betanews, on their newly acquired "physical platform" which is oh so subtley mentioned in the header of every graph, can obtain these numbers on a dedicated platform does not reflect real world numbers.

Do you take into account that most people who read this site fall into, typically, one of two categories. Those who use their computer to its potential and aren't afraid to have 70 processes running at once to make your daily life easier and those who install very few applications in an effort to minimize memory usage and try and obtain a "faster system". With that said, there are always exceptions but from reading reviews and posts on this site, I have come to that conclusion and I find it fairly accurate. With that said as well, what applications are running on the betanews test machine? Is it dedicated to producing these browser graphs? If so, then that is far from accurate to real world. Is it a dedicated test machine for every aspect of the software world? Again, a possible lack of accuracy to real world. The problem, again, is that you can't accurately produce results which will give you a viewpoint of any one area of the world when it comes to browsing.

Browsing is a very personal thing and everyone has their own habits. If betanews test included simply going to a site by entering its URL and hitting GO or ENTER then that is one method of testing. If you do it by clicking a link from another page which might not be finished loading, that is another way.

Case and point is that I feel betanews efforts could be better spent on REAL SOFTWARE NEWS than worrying about theoretical numbers and percentages which at the end of the day do not equate to anything near what the average user could expect to see.

My two cents.

Yeah, the number comparisons really are pointless. People will use what they feel most comfortable with and what suits their preferences. Any loading and execution time currently saved by switching to another browser would be lost during the process of installing and learning the new software.

Having said that, as web apps become more popular, execution speed will be very important (though I'm not convinced JavaScript is the way of the future). So I don't mind that sites like Betanews are growing more focused on benchmarks.

No, I won't call you stupid, FixXxeR, because it's a fair question. I'll treat it fairly, I promise.

Why is speed suddenly so important now? When the browser battle between IE and Firefox started becoming...legitimate, the manufacturers began using real-world characteristics to distinguish what makes them better than each other. Both sides started using speed as a determining factor, which awakened some long dormant impulses among those of us who remember when speed mattered in programming.

Then Mozilla started really turning up the heat, this time not just in the marketing department. The onset of TraceMonkey, the organization's new JavaScript engine, made folks start to think seriously about why speed should matter in the first place. It was "game on" at that point.

And then Chrome came out of nowhere and started blowing things away in the speed department; and granted, even now it doesn't have all its belts and bolts on straight, but as an engine it's something to be admired. Now, I'm a speed junkie from way, way back, and this was when I decided it was time to pay attention to speed again. It was about that time that Safari strapped on a couple of warp engines and began taking this seriously too.

Now, your other big question here was, would a thousandth of a millisecond here and there make any difference as to whether you'll use an application? My answer there is, yes, but not for the reasons you might think.

First of all, I think it's just a given that a well-tuned application is a better application. If it's faster, leaner, and more efficient, it is less liable to cause problems, it will be easier to maintain, and folks will like it better. Speed is an indication of quality. If somebody cares enough to tweak a program to squeeze that thousandth of a millisecond of efficiency out of nowhere, he cares about the dynamics and the architecture of the application at a level that most developers nowadays have completely forgotten. I remember when I used to time competing word processors' global search-and-replace functions on a stopwatch, and I remember the day I discovered that Microsoft Word became faster in numerous departments than WordPerfect. That was a day I sent up a green flag and said this race was on. And today we know the results all too well -- speed to work and speed to market were the two reasons for WordPerfect's demise.

When developers care enough to make speed matter again, they care about quality again. And that's why I'm all over this.

-SF "An-n-n-n-n-nd He-e-e-e-e-e-e's ON IT!" 3

This is not about the OSes or the OS performance, as the author would like to claim.

This is more reflective of the developers of each product and how well they understand the platform.

For example, if you are a 'bad developer' you might allocate more memory to do 'in application' caching or use APIs and API tricks you learned back in 1992 that are no longer even needed and are slow in comparison to the APIs that replaced them.

So this test actually shows which set of developers have a 'clue' about OS optimizations and the proper techniques to get the most of out the OS they are running on.

Sadly, most developers still have NO CLUE about what Vista or Win7 is doing, so any 'help/performance assists' that Vista and Win7 brings to the table is lost on these developers and their projects.

And this list of things developers don't get go all the way from dealing with Superfetch and 'stop caching large amounts of data that the OS should be doing for you' to even using the proper API calls.

Take the Game developer that sucks all 2GB of your RAM to cache textures that Vista or Win7 will learn to do automatically after a couple of runs of your application, and they are creating more work for themselves, and can't cache as efficiently as the OS. (Games that have 'in game cache options' if you set their cache down to nothing on Vista or Win7, they run faster, as the OS does do a better job after you run the game a couple of times. And the RAM the Game is sucking up for 'cache' is better used by the Game or Vista/Win7 using it to prefetch/superfetch/cache for the Game.

Also take the developers using outdated APIs or methods that go the long route when a simple API is available to do what they took 50 lines of code to try to do themselves. When a new OS offers new API features, maybe people should pay attention, it can make a big difference.

Bascially these tests show you what development teams 'get' Vista and Win7 and what developers don't or are still using outdated coding practices.

PERIOD.

Well, Anthony, you're kind of right...The cross-platform comparison speaks to both departments. First, you can see how some Web browsers are generally faster in XP than elsewhere, and that gives you a general benchmark. There's less overhead in XP -- doesn't make it a better operating system, in fact, because that overhead in Vista and Win7 is also in the security department.

But once you take that benchmark into account, you can see how some developers probably have optimized for different platforms. I suspect Apple's Safari developers had XP as their principal Windows test bed -- Safari 4 is a lot faster in XP, and during the beta process, at one point, the Win7 version was slower than the Vista version. And that could be an indication of precisely what you're saying -- that developers may not have a clue about what's over their heads until they run into it.

-SF "Oof!" 3

When quoting these numbers, you really need to mention the margin of error. For example, the Opera 10 beta results quoted above are probably within the margin of error, and so Win7 and XP scores are statistically identical.

I've said it before and here it is again: Too many variables in web speed to lose sleep over it and even discuss it. Weather, distance, node users, cable company, the site itself, shall I go on? If no one can wait milliseconds for a page to load then you've got a problem. I have had times when FF is slow as Hell, and times when it is blazing fast, so these speed tests don't mean s*** to me and never will. One thing they do though is bring in ad money for the web site. Don't mean to sound as if I am flaming everyone, but for Christ's Sake, get a grip already. If you are on this site then I presume you are a power user or at least well learned in computers, so why are you worried about milliseconds and nano seconds? As far as the OS, I have seen no significant speed increase in browsers in W7 or XP, at least not enuff to shoot off some fireworks. Keep you OS tweaked and clean of all junk and I guarantee that you will have no issues with speed.

I love this kind of stuff. With me it used to be how fast my car could get to 60 MPH from a standing start, but since they've seen fit to take my licence away I now have to settle for having the fastest browser. But having installed all known browsers and compared them I'm puzzled. Those who lead the field on these tests may be the fastest on the tests, but in the real world IE8 is proving the fastest on my machine. Chrome now has it's ups and downs, sometimes fast and sometimes slow, go figure. Opera, well Opera is Opera, sometime OK and sometimes pretty ordinary.Firefox has become high maintenance and slow alas. So I see these tests a bit like the stuff you read in your car mags, that would be the 0 to 60 on a straight road, with no traffic in the way, no bends, no bumps and no police cars. Trouble is those pesky bends and pesky police cars are always slowing us down, so until I get my licence back I'll stick with IE8 until the next batch of tests arrive.

Now only if they fix the broken Windows Explorer in Windows 7, I'd gladly upgrade without looking back again.

This is a somewhat useless comparison between XP and Windows 7. I loved XP and recently upgraded to Windows 7 RC (skipped Vista). Windows 7 is TONS faster, even with Firefox and here is the reason why...64-bit!!! Who runs XP 64-bit...I'd say very few. I sure wasn't. By the time I knew what 64-bit was (just before Vista was released) and bought a 64-bit CPU I wasn't going to shell out all that money for XP 64. But when Win 7RC came out I jumped at the chance to go to 64-bit and WOW!! Huge performance jump! Then I upgraded to 6GB RAM (not possible with 32-bit) and WOW again! I'll definitely be purchasing Win7 64-bit when it is released!

So, while I don't dispute the accuracy of these statistics, the practical reality is that anyone who is on XP has hardware that supports 64-bit (or buying a new computer) should go with Win 7 64 or Vista 64 (with free upgrade to 7 starting June 26). And if you are that rare bird who has XP 64 then more power to you and ignore this advice.

Wake up and smell the coffee. XP is still installed on 75% of the all the PCs in the world.

Installed yes, but it's still old and outdated.

The browser speed differences are so minimal that the decision on which browser to use should come down to personal preference. The OS doesn't even come into play for me, I use Firefox for all it's functionality. If I need IE for any reason, the URL gets entered into IE Tab's sites filter. Someday soon the lag time issue will be fixed since Win 7 & Firefox 3.5 are both betas still.

Bloated this, bloated that. You ppl don't even know what bloated really means :rolleyes:.
These tests are also funny. Chrome and Safari always ahead of everyone else. While in reality, they are faaaar behind. Why? Firefox an Opera don't have to wait for s***ty slow ad/banner servers that always take forever to fetch huge flash banners and other junk. Thanks to available AdBlock feature. Any real world page that i tried viewing in these two opened way faster than in Chrome/Safari, despite these tests showing exactly the opposite. That's why i still think these pure JavaScript tests are stupid and useless because there is no web page that consists entirely on JS and does something this demanding to show any real world difference between all known browsers. I also don't see any real difference in GMail which is suppose to use JS most intensively.

No kidding it runs slower on Vista SP2.5. Vista is a bloated pig.

Vista SP2 ran just fine on my 1 1/2 year-old PC. Anyone still making this old and belated argument either hasn't used it since it's beta days or is still using it on an outdated PC, and if that's the case you should be upgrading to Windows 7 anyway.

Does this surprise anyone? Why don't they test it in Windows 95? I'm sure it will be even faster. Windows 7 is doing more then Windows XP is. I even doubt anyone would see a difference between the two unless they run some benchmark.

Some would argue that it's DRM bloat in Vista/7 that slows it down so much.

OSs don't necessarily need to get slower and have more features for new versions. It's perfectly acceptable for an OS to aim to be slim, trim, and faster with each revision. Ubuntu is going for that I believe. (10.4 is aiming for 10 second boot time).

Thats just total BS. you have no idea what you are talking about. Windows 7 is not bloated, it's in fact more trimed down and faster then Vista. and most fuctional and powerful then XP. So learn to like it. It's a real good OS.

That's what we heard from Vista apologists when it came out: "why are you surprised that it is slower than XP, it has new features". Well guess what, Windows 7 has even more new features yet it is much faster and slimmer then Vista. The point is, when Microsoft was pushed (by people not accepting Vista) it turned out that IT IS possible to add new features without turning OS into a slow pig. Microsoft obviously did much better job in that regard with 7 than with Vista, but still not good enough compared to XP.

About Windows XP being slower then Windows 95, XP compared to 95 had architectural changes like protected process address space which made it more stable but slower. Vista/7 have no architectural changes mandating slower speed, so they are slower for no reason at all (other than sloppy design/development practices).

Why in the graph the browser "Google Chrome 3" not is showed as beta version?

They should just call all of it beta. There isn't a month that goes by that a stability or security bug isn't found with *any* of these browsers.