Microsoft claims Firefox- and Chrome-whopping IE8 speeds Microsoft is touting its own research that claims Internet Explorer 8 - expected next week - is faster than Firefox 3.05 and Google's Chrome 1.0. The company has produced a 14-page report (pdf) and supporting video that claim IE8 is three times as fast as Firefox when it comes to loading web pages and one third quicker than …
I wonder how much of that extra speed is due to the use of secret APIs that bypass inefficient Windows kernel functions, and open ever wider security holes by in effect making IE part of the OS.
MS never has seemed to understand what modularity is about, if gossip can be believed about Excel and other apps having secret trapdoors into the guts of the OS, and parameter lists being insanely long for the APIs MS has made public.
Thats the funniest thing i have read all day, almost as amusing as Ballmers admission of 'Challenges' with IE8 and his statement that it is apparent that websites are not yet ready for our software (IE8). Now forgive me but the website were here first you conceited tossers. shurely shome mishtake - Ed
A browser's performance depends on more than just itself. For instance browser performance when in use depends on what it's connected to, the PC it's installed on, the OS it's running, the speed of your firewall/AV software, and how much 'crap' is installed on the browser.
Whilst it's encouraging to hear that Microsoft is now paying more and more attention to the performance of it's bloatware, I find myself strangely unimpressed by this particular anouncement. A bug-free and secure browser is worth more personally.
I quite like IE... but that is rich!
Chrome leaves it for dust. Safari Beta 4 has taken some nice tokens out of Googles book, and the cover flow history is awesome, but still performance isn't good.
Is this some sort of early april fools. They couldn't keep a gag like that till April (Holly, Red Dwarf).
You'd have thought so, wouldn't you? Page loading speed is, however, dependent largely upon how many parallel connections your browser will run to download any images, CSS and javascript files on a page. Another bottleneck is any client-side, ie. Javascript rendering especially AJAX which is why Javascript has been getting so much attention over the last couple of years.
At the end of the day it's nice to see Microsoft taking internet browsing seriously again.
Sounds great! Where can I grab a DMG for my home MacBook? Not available? What about the source code so that I can build it for our office Linux desktops? Not available either?
In the end, who really cares about a closed-source application that only runs on a single, closed-source operating system?
"Finally, Microsoft said it measured page load times using "visual cues" or a combination of visual cues and the ability to interact with a page to determine if a page had completely loaded."
That'll be the same kind of visual clues that the windows shell (desktop) throws up around 2-3 minutes before its capable of doing anything meaningful.
I'm actually looking forward to IE8 if it really does make inroads into supporting any form of web standards rather than the in-bred crazed rendering scheme that riddles everything prior to it. Unfortunately MS's idea of optimisation has always been around throwing more and better hardware at the software until it runs "adequately". Hopefully with more competition and with more awareness of other browsers they'll all improve.
But then, MS's aim is to make everything "require" custom ActiveX (or .net / com as they've decided to re-brand them today) add-ons and through that ensure that everybody is forced to use windows and IE. I'm just waiting for the first critical, or damn hard to do without, parts of windows, to require the abortion that is silverlight.
I'm sure they're bang on with IE 8 being faster, but if you're like me you prefer your browser to show more than a few script errors when you visit a website.
What I always wonder is why Microsoft programmers don't like making software that's compatible with the standards everyone else uses. Is it some kind of arrogance? Do they really think that everyone should re-produce their websites to make them compatible with Microsoft browsers.
"I would have the thought the bottleneck on my 2-core AMD running at 2.7Ghz would be the internet pipe"
Indeed. As the article notes, the tests were done on cached content, so we'd be talking about LAN latency, if not hard-drive or RAM latency. That's probably 3 or 4 orders of magnitude faster than any real-world scenario.
For a realistic test, trying pulling the content over a few megabits of broadband, like the rest of us. If your browser cannot render *as fast as the pipe delivers* then that browser is irredeemably broken. It shouldn't even be possible to compare browsers, since a percentage point of variation in the pipe should overwhelm any difference elsewhere in the chain.
This "test" is just benchmarketing.
Hmmm... lets see?
They measure the time it takes to render a web page to lets say for arguments sake 1 millisecond, but, and its a big but (fnnarr), Firefox takes 3 milliseconds. so they're three times as fast.
...well yeah, but the page takes 10 seconds to download, so IE 8 serves it a teensy-weensy bit quicker then Firefox ! Gosh, I'm sure I'll notice that over a session.
Wow Microsoft, some very clever programming there! ~ or perhaps as Rob points out above, you've just bypasssed any sanity checks (purely in the name of research of course).
When will M$ realise that what we want is not, necessarily, fast software, or bloated eye candy? just honest reliable software with, perhaps, the option to turn off some of the more commercial ly minded features.
I know you want to me to sell my soul Mr Balmer, but I'll only do it if the price is right.
yours
Uncle Sid.
So, if you strip it down and use some kind of undisclosed subjective criteria, IE8 is faster at loading pages from its own cache. I guess they turned off all these pesky security add-ins too (you know, the stuff that is build in every other browser). Yay. This probably means that in a real-world configuration, and using objective criteria, IE8 is still wayyyy slower than any of its competitors. While massively less secure. MS are really desperate aren't they?
I guess they measure a page as being loaded in the same way as they proclaim Windows is usable as soon as the desktop appears. It's likely not usable or even scrollable for another couple of minutes - per page.
The first hint of something awry with MS measurement is non-adherence to any accepted standard (including, but not limited to the El Reg measurements). The second hint (not applicable here but in general in an MS presentation) is an absence of source quoting or methodology used. An example: Windows vs. anything-but-MS virus infections (just in case you forgot that trivial benchmark in the fight to keep your information private)..
The average PC user (and by PC I mean Windows, Mac, Linux, Unix, etc.) really WILL NOT notice or even care - unless SOMEONE makes a big deal about "fast" and they start checking.
In the "real world", I've found IE(5-6-7) to be similar in "speed" to FireFox or Opera. "Real world" is sitting at your desk in a large corporation bringing up internal web pages on your 1MHz Dell or Compaq cheapbox while running Word, PowerPoint and Excel along with all the other apps you need to get work done. Or at home, connected to the Internet via a 1.5Mbps DSL link, with all your Facebook, chat, Twitter, camera, music and video software pounding away.
Simple case in point: surfing the internet at my inlaws I was connected via a SLOW WiFi link while my sister-in-law was cable-connected. I have an OLD 1.4GHz laptop, she has a brand new MacBook Pro with a 2.4GHz dual-core processor. My pages loaded instantly, hers were taking MINUTES to load. However, she had 25 apps open, including music, Office, chat, mail, etc. and I had only the browser and email.
When I asked her why the Mac was so slow, her response was "It's not slow. This is the way it ALWAYS runs!" Oh, and all 25 apps are automatically started every time she powers up.
I rest my case.
to me means one thing.... someone sitting there with a stopwatch in their sweaty palm.
Or, given that this approach was to be taken, who's to say that IE 8 doesn't trigger visual cues in anticipation of completion, rather than as the very last thing performed in response to actual completion?
Either way, this is hardly a solid basis for drawing conclusions from results where differences often come down to 100th's of a second.
Then again, I haven't read the "methodology" paper....
Maybe Microsoft really did give themselves a hard time and were pleasantly surprised to find that their own tests of their own product put it out front.
"I wonder why they didn't compare it against Safari beta 4 or Opera?"
1) They compared against shipping versions so beta's don't qualify
2) Or Linx, or "insert your webbrowser here..." - Do your own test!
The real issue for me is who really gives a f* about 0.2 seconds of difference in loading a page!
btw. If you do watch that video - it's only a 47MB download - which is mostly text and a simple tune!
from an untrustworthy company.
there are so many factors that determine the outcome of such tests that i don't think ms are being completely honest. line speed/traffic, what sites etc.
after a while will no doubt slow to a crawl anyways like most ms software.
i couldn't be bothered because if IE8 has a 'compatable for some "approved" sites list' i won't ever be going near it.
oh, and HAHAHAHAHAHA. funny stories and comments today.
Faster?
Even if these figures are legit, if you look closer you find they're admitted to be figures for loading pages from the cache NOT for actually USING a website.
Since IE's JavaScript performance is so abysmal, it's going to be dog-slow actually using something like Gmail. Chrome, Safari - and even Firefox - absolutely destroy IE on JavaScript performance:
http://news.cnet.com/safari-challenges-chrome-on-web-app-speed/
Microsoft knows this. IIRC, Google was even advising IE-using customers to use an alternative browser if they wanted reasonable performance with their web apps. Microsoft doesn't want to be known as the company with the dog-slow browser - but the public's starting to realize that that's exactly what it is. THAT'S what this deceptive set of figures is about. It's tantamount to lying - lying by omission and distraction.
Their report says IE8 loads google.com 0.02 seconds faster than Firefox. They are "measuring" this by recording 30fps video and then watching it back - so there will be accuracy of no more than a frame, i.e 0.03 seconds - and from the introduction it sounds like they also performed the test a grand total of 2 times. Awesome. No margin of error on the figures either.
Oh, and they completely reimage machines between tests. Just like users do before they use the internet...
As the test parameters appear to be completely subjective, this is a load of carp. How about testing from the time the http stream starts to the time the http stream stops?
No? Why? Because they know that they can't get better performance than ANY other browser out there.
Mine's the one with the copy of Lynx on a USB drive in the pocket.
@Wun Hung Lo: "They must have a new definition of "faster"."
MS is well known for redefining terms when it suits them. For example:
"W7 will allow you to uninstall IE8"
actually translates to
"W7 will allow you to remove the GUI front-end to IE8, but the engine and the guts of IE8 (including bugs and security holes) will still be there and used by W7 for other purposes"
When I uninstall something, I want *all* of it to go - not just the user interface.
Pissing me off, because of all the page errors, secured site false alarms and other little graphical problems?
FU MS!
What I've experienced on all of the ie8 releases is that it sucks worse than ie7. I'll stick with IE6 when I'm required to use an MS browser, otherwise it's Firefox for me, baby!
If you go purely for "Visual clues", then Vista does indeed boot in ten seconds. However, having lately had this monstrosity foisted on me, I can tell you that the thing is not usable until the little drive light stops blinking, which takes a few minutes, or one trip to the coffee machine.
Paris, because she's all about visual clues.
*BEST*
*METAPHOR*
*EVER*
Here, have an e-beer!
Who gives a shit what M$ claims about their "software"? Anybody with the slightest hint of a clue knows to stay well away from it, and the rest deserve to suffer. If they want to claim that their "software" is faster, or that they worked out the square root of negative one, then fine - Doesn't mean I'm listening...
"In the end, who really cares about a closed-source application that only runs on a single, closed-source operating system?"
Your mum, your dad, the person down the street, your postman, your milkman, the guy who sells you food at the shops, the random person you bumped into in the street (this is of course assuming you actually go outside) basically most people in the world really don't care that they can't get access to the source code.
Oh and most businesses like it cause its all nicely integreated in general.
Stop complaining about ms and do something if you actually don't like, instead of sitting in your rooms/office bitching because you have no idea how a market works or what a business plan is.
'Where can I grab a DMG for my home MacBook?'
And later in the same post:
'In the end, who really cares about a closed-source application that only runs on a single, closed-source operating system?'
Anyone else notice the irony apparent in having these two statements in the same post?
I hate MS with a passion, but when it comes to creating closed-source proprietary technology designed only for their preferred platform Apple are surely the undisputed world champions?
'If you go purely for "Visual clues", then Vista does indeed boot in ten seconds. However, having lately had this monstrosity foisted on me, I can tell you that the thing is not usable until the little drive light stops blinking, which takes a few minutes, or one trip to the coffee machine.'
That's just because it is a new installation Monstor. Give it a couple of months and you will be able to have fourteen cups of coffee, read a medium-sized paperback and learn a foreign language in the time it takes to boot.
I suspect that the other posters have identified the key phrase -- "visual clues".
Microsoft has optimized the engine to give priority to reading on-screen images from cache and running scripts that impact the layout of the part of the page being rendered.
There are many ways that this might be tied into special OS APIs. One way is using asynchronous I/O with priorities. Another is giving storage layout hints when initially caching e.g. grouping objects together on the same disk cylinders (hmmm, which would give the best results with a fresh install.. coincidence?)
If that's what Microsoft is doing, they are making a traditional Microsoft move. Tying an application tightly to an optimized (but private and obscure) OS interface looks great for a few years. Then third-party applications figure out how to use it, just as it's time to abandon the interface to let the OS design move forward.
It's possible that the IE development team isn't doing that. They might just be optimizing the display engine to prioritize rendering the visible part of script-based web pages faster, without using special OS APIs. Pretty much as interlaced GIF images could be displayed with low resolution before finishing a page load over a modem 15 years ago.
But my bet is on funky OS APIs.
Downloaded the new IE 8 and initialized it with most options turned off - the plain browser. Got the biggest web page to which I have ready access - about 6.6MB - in a disk file, loaded it in both IE8 and FF3.07, then timed refreshing the page in Offline mode (neither should try to look to the net). FF3.07 took 2.3 seconds, IE8 took 3.4 (average time over few runs to reduce effect of operator timing error). FF never took more than 2.5 seconds, IE8 never less than 3.2 seconds. I like to keep IE on my machine to check the occasional page the FF seems to be handling strangely - but otherwise use FF all the time.
Funky APIs... or, given that it's all based on subjective "visual clues" when loading *cached* pages... I bet the easiest way to do it would be to snapshot pages in the cache, and display this pic while waiting for the real page to load. Tadaaa, instant loading! I mean, if you're gonna implement a borked browser, might as well go all the way down.
PS if someone is dishonest enough to really use this method, I want a cut.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
