Slashdot Mirror
Google Frame Benchmarks 9x Faster than IE8
ChiefMonkeyGrinder writes "Early tests with Google's Chrome Frame found IE8 runs 9.6 times faster than usual. The testers ran the SunSpider JavaScript benchmark suite." The other question is what is the performance hit of using the Frame plug-in instead of running the browser natively.
However it seems like they only measured JavaScript engine, which by no means contribute everything to how fast browser or browsing feels. And everyone probably knew already that Google's JavaScript engine outperforms MS's (and being one of the main thing Google's thing use, they have a reason to optimize it till its dead)
This seems to be the usual thing with other browser benchmarks too, they only benchmark the javascript engines and similar under the hood things. Yeah it's easier, but it doesn't really tell the truth.
User interactions and GUI responsiveness contribute a lot, actually even more so, to how fast browsing feels. IE is horrible with this and has always been; everything lacks behind, scrolling is galaxies far from smooth and the general feeling is just bad. On that note, Firefox suffers a bit from the same things. I think only Opera and Chrome have done UI responsiveness good. Which also brings the question, does Chrome Frame improve it on IE too?
Looks like Google are going to try and beat Microsoft at their own game:
Embrace, Extend, Extinguish.
So, Google Frame upgrades the engines...on the Titanic?
Google doing this speaks a lot to the character and principles of their company...that is if you trust big companies. I'm not surprised to see that IE 8 is running faster on the Chrome framework. All my experience with IE 8 confirms why I don't use IE. It has been very unresponsive for me in multiple situations. I'm sure this is one of many steps Google is pushing for to "speed up the web".
From the article: "Notably, IE8's SunSpider scores with Chrome Frame running equaled Google's Chrome browser, a solid indication that the plug-in effectively turns any version of IE into the speed equivalent of Chrome itself." So no, that isn't the "other question".
You're wrong there mate. On our corporate intranet there's a section of javascript that's 256k in size. IE6 (corporate standard) takes about 20 seconds to load that while Firefox loads it instantly. It's not about how fast the Javascript is received, it's about how fast it's rendered.
...what's the ACID3 results for such a combo?
Quo usque tandem abutere, Nimbus, patientia nostra?
It's no wonder Microsoft is claiming that Chrome makes IE less secure. If it lets IE run eight times faster that means that there will be eight times the rate of security breaches. Oh Noes!
When our name is on the back of your car, we're behind you all the way!
The other question is what is the performance hit of using the frame plug-in instead of running the browser natively.
FTFA: "Notably, IE8's SunSpider scores with Chrome Frame running equaled Google's Chrome browser"
Microsoft has issued the following PSA: 'Some users have been found to experience sides effects from a sort of 'digital whiplash' after installing the new Chrome Frame plugin for IE8. This is not a risk we would recommend our friends and families take.'
Caffeine is my anti-drug!
Duranin - A NWN2 Roleplaying Persistent World
I like the idea of segregating the browser's user interface/menus/controls skin from its rendering engine and plugin-model guts. Even better if there was a standard plugin API across browsers.
I would love to be able to pick any "shell" and put in one or more "guts," and even flip between them on a per-URL or per-site basis.
Of course, the security risks of this are not small. But still, it would be cool.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
Likewise, I've seen javascript which manipulates large datasets, which takes the lion share of time to run; somewhere in the 30-60 second range. Recent javascript performance boosts have allows such manipulation of large datasets to become feasible and even practical.
The truth is, more and more people are attempting to use a browser as a general purpose user interface for many applications which were previously considered unattainable with older browser technology and I only see additional momentum building in this direction.
Fast rendering and javascript is a make or break for most of these types of applications.
Improvements and optimizations like this are ALWAYS welcome from any quarter.
Sure, for a single page - w0000t, i saved 0.1 seconds doesn't do it many favours, but when you consider:
lots of users open multiple tabs and load many hundreds of pages a day those savings start to add up.
Additionally, any optimizations done usually go towards powersaving and extending battery life.
in a mobile embedded platform its vital (nokia maemo platform for me..) and I'm sure these improvements and benefits will trickle down eventually.
well done google, keep up the good work.
liqbase
Google is building and expanding its own online infrastructure (see weave for example).
It's also undeniable that IE has a big market share, and Google needs to account for it. So this seems to be the only reasonable route they can take as this issue doesn't seem to be of much interest to MS at the moment.
While IE is our "company standard", they don't care if you prefer to use another browser.
However! Most of our corporate intranet applications will ONLY work on IE.
( *cough poorly written proprietary crap cough*)
So now with Chrome infecting my IE, I have no way to access vital corporate apps.
There is only one type of consumer who should be interested in this: corporate users who do not need IE for specific webapps, and whose companies will not let them install other browsers, yet will let them install plugins.
How many of them can there be??
"Notably, IE8's SunSpider scores with Chrome Frame running equaled Google's Chrome browser, a solid indication that the plug-in effectively turns any version of IE into the speed equivalent of Chrome itself."
Last paragraph
In the real world its never the browser we're waiting on anyway. it's the connection or server on the other end that we're always waiting on.
I've seen Slashdot's home page freeze for a minute on Internet Channel on my Wii. I don't know whether it was a reflow or a JavaScript, but it was frozen.
I'm going to go out on a limb here by /. standards, and say that this is a very good idea that is a neat technical solution to a problem. Google's goal is simple : their core strength is that they are incredibly good at creating and hosting web applications. They have some of the most reliable and least expensive (per unit performance) data centers in the world, and they have some top notch coders that have created some amazing applications. The problem is that web applications have to run in web browsers, 20 or more layers of code away from the processor on the host. There's unbelievable performance slowdowns compared to a native application. Speeding up the browser would make many google applications more responsive and compelling, and google could care less whose browser it is. They are freely licensing the chrome code for inclusion in other browsers.
The problem with Chrome is twofold :
1. It's an unbelievably complex task to make a web browser work with every website. Mozilla and the Microsoft browser team have hundreds of developers that have worked for years on their browsers.
2. It's very difficult (and expensive) to get people to change browser. Microsoft wins by default most of the time.
This browser plug-in solves both problems. Now, only websites that the developer knows will render properly in chrome will call on the plug-in. Users will continue to use IE8, oblivious to the fact that some websites are actually being displayed using the chrome browser engine. Google applications will of course all properly render in chrome, and they will be set up to encourage you to download the plugin if you're running internet explorer. Some google apps may even require it, much like you need flash to see youtube videos.
The only problem with the approach is overhead : obviously keeping multiple browser rendering engines running at the same time will eat up a hundred extra megabytes of memory or so. You know, about $3 worth of DRAM.
I'd add that this isn't the first time this sort of thing has happened some later versions of the Netscape (version 8?) browser implemented a compatibility mode that ran IE inside the content area of the browser instead of Gecko for certain sites. A fat lot of good it did them too.
I've had Slashdot freeze on Google Chrome all the time, I still have no idea why.
Don't forget, installing satan's chrome frame means your children and all their family will suffer for all eternity, in HELL!
For some reason yet undisclosed by MS.
If Google were to do something like this with Safari, would Apple allow it? Or will the next update break it? (I know both are based on webkit, and Safari doesn't need the feature and speed boost, but just wondering out aloud).
This space for rent.
I'm going to go out on a limb here by /. standards, and say that this is a very good idea that is a neat technical solution to a problem. Google's goal is simple : their core strength is that they are incredibly good at creating and hosting web applications. They have some of the most reliable and least expensive (per unit performance) data centers in the world, and they have some top notch coders that have created some amazing applications. The problem is that web applications have to run in web browsers, 20 or more layers of code away from the processor on the host. There's unbelievable performance slowdowns compared to a native application. Speeding up the browser would make many google applications more responsive and compelling, and google could care less whose browser it is. They are freely licensing the chrome code for inclusion in other browsers.
The problem with Chrome is twofold :
1. It's an unbelievably complex task to make a web browser work with every website. Mozilla and the Microsoft browser team have hundreds of developers that have worked for years on their browsers.
... does it have adblock yet?
If only I could download Google Frame...
Gotta love synthetic benchmarks!
Looks like someone took this seriously!
I iz in ur browzr, fixn ur renderer
# (/.);;
- : float -> float -> float =
I'd argue that the real metric is user perception, not raw performance. If you can keep the user entertained with a screenful of flying monkeys, you can actually load slower but be perceived as faster than a blank-to-full-screen transition.
John
The truth is, more and more people are attempting to use a browser as a general purpose user interface for many applications which were previously considered unattainable with older browser technology and I only see additional momentum building in this direction.
We are doing *exactly* this. We provide a hosted, vertical software system, and for years we've done everything in our servers.
However, recent builds of the FireFox JS engine are fast enough that we can start moving the processing out from our hosted application server cluster into the user's browser. The users love the results - applications that load in a few seconds, and run from their computer at near-native speeds, accessible anywhere.
But, rather than spend inordinate amounts of time trying to get stuff to work in IE, we simply require Firefox. That way, we can support Windows, Macintosh, Linux, and any other platform that runs FF 3.0+. It's not been hard for us to make this requirement, basically only minor complaining from techs.
Our customers are more interested in "Cross Platform" meaning "Can I get it to run on MY computer" than "Can I get it to run in MY browser".
The evolution of javascript performance is an industry changer - it's what makes hosted applications actually WORK, despite all its warts.
I have no problem with your religion until you decide it's reason to deprive others of the truth.
Making this installable and usable under Win2k would go a long way to getting people to facilitate the move away from IE6. Firefox works but tends to slow down on lower end hardware. I tried it today and it didn't install. Maybe there's a way to make it work by manually copying files.
You can lead a man with reason but you can't make him think.
The key piece missing is standardization, so I can mix and match.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
The other question is what is the performance hit of using the Frame plug-in instead of running the browser natively.
Well let me give you a hint: the native browser renderer is a plug-in itself (well known as mshtml.dll). The actual other question is, what is the point of this plug-in in reality. People who use IE can install Chrome today already. Those who keep IE mostly do it for two reasons: 1) it's a corporate policy and their business apps need it, and 2) they don't know any better. So the frame addresses none of those two segments adequately, since Google Frame is not 100% compatible with the standard MSIE stack, and requires people to deliberately install something (at which point they could as well install Firefox or Chrome itself). The only dividents to be had from this project appear to be political. Last time some competition showed up for IE, Microsoft put vast resources in order to catch up and create IE7 and 8. The prospect that their competitors will take over IE piece by piece via parasitic plug-ins must seem even scarier from their point of view.
Give us some benchmarks for IE6. I can't think of a single reason for anyone to run this on IE7 or IE8.
Populus vult decipi, ergo decipiatur...
"Force shits upon Reason's back." - Poor Richard's Almanac
Do you really not allow WebKit? I assume they've given thought to this ... "anything but IE, you deluded fool" should be enough. Or "I see you're using IE. Get the Chrome plugin or a real browser, you deluded fool."
http://rocknerd.co.uk
"Google Frame Benchmarks 10x Faster than IE8" is a more accurate headline, since no sane rounding scheme in the world would round 9.6 into 9.
GAAH! MY PRINTER IS ON FIRE!!! PUT IT OUT! PUT IT OUT!
I found the original article, but it still didn't have the numbers from the test. What it does have is a bar graph jpeg of the results. So I measured them, and the two scores are 24 pixels and 232 pixels. 232/24=9.68, which is close to that 9.6 number they're giving.
But, they were saying it was 9.6 times faster. That is wrong. It is 9.6 times as fast, or 8.6 times faster. It bugs me when people get that wrong.
like the rest of /.-ers.
With FF (Linux and Windows), with heavily commented on article, I often get the message that says JS is taking too long to load, would you like to stop/continue loading it? The only sane choice is to stop. In that case, most of the comments become buried and I'll have to RTFA to know what's going on.
The only possible interpretation of any research whatever in the 'social sciences' is: some do, some don't
I thought we all use Linux here?
Oh yeah videos in a stupid proprietary format, we all need more of those...If you argued for OGG Theora embedded HTML5-style I could agree.
"When information is power, privacy is freedom" - Jah-Wren Ryel
They ran the SunSpider benchmark. Sadly, it only tests certain parts of JavaScript, and is intended to test whether the JS engine is optimized for the CPU it's running on. But just because your JS engine is optimized for the CPU doesn't mean that the whole browser is faster or that it renders pages faster. I'm sure it actually does, but this test does not show that at all. I wish people would stop blindly repeating Apple's marketing nonsense about SunSpider.
Clever signature text goes here.