Slashdot Mirror
Benchmark Battle, September 2015: Chrome Vs. Firefox Vs. Edge
An anonymous reader writes: The next browser battle is upon us. Edge has been out for more than a month, and its two biggest competitors have received significant updates: Chrome 45 and Firefox 40. This article puts all three through their paces, and each manages to win a few tests. Edge convincingly won the JetSteam and SunSpider JavaScript benchmarks, while also eking out a victory in Google's Octane test. Chrome was victorious in Mozilla's Kraken benchmark for JavaScript performance, while also edging out Firefox in HTML5Test and the Oort Online WebGL test. Firefox won the WebXPRT test that combines HTML5 and JavaScript performance, and also the Peacekeeper test for general browser performance. There's no clear dominant browser for performance, and none of the three are obvious laggards, either. Browser competition seems to be in a good place right now.
Turning off JavaScript makes baby webdeveloper cry.
With FF and Chrome JavaScript performance has been good enough for most web apps for the past couple of years. Good to see IE has caught up, but JS performance is not been what's hold the web back. The issue is DOM performance. For example, my stock trading app eats up 100% of my CPU time and causes Firefox UI to get unresponsive when tracking more than 30 stocks. Their native IOS app barely uses any CPU time and runs better on my phone than the web app on my desktop. It's time we ditch DOM's document based model for something application centrist.
Several of these benchmarks are outdated to the point of barely meaning anything anymore, chiefly Kraken and Sunspider. But even they have value compared to HTML5Test, which artificially skews data in favor of Google-specific technology, and breaks the scores down in mysterious ways to make Chrome's lead look much larger than it really is. At least go to caniuse.com and tally up the support there instead. Its tests are far more accurate, cover FAR more features, and aren't artificially biased.
Performance is roughly the same. Chrome renders HTML5 more correctly than others.
Tests were performed only on a Windows system. Still a very valid test as that what most people use.
I will give up 1% winning speed, or 50% actually, for a browser whose various scripting engines don't grind things to a halt as some web site overwhelms their calculations, no doubt abusing this to grind defenses to a halt, or exploit race conditions.
Google, with the "best" programmers, this means Chrome! Some damned 3D ad or video ad or something, freaking stop it from "use every CPU cycle you can!" to eke out that winning percent. Maybe they overuse RAM, too? You figure it out if you are so good.
(-1: Post disagrees with my already-settled worldview) is not a valid mod option.
It still made a really good showing by winning 3 (tied w/ Chrome) and the ones that it wasn't 1st or 2nd in seemed to be related to lack of codecs which will likely get updates soon since there was an article about them adding support for VP9 coming up.
Really? I have never noticed that anywhere.
What it normally does is break website menus.
Minimum threshold fixed. Thanks!
What happens when you turn off Javascript?
God kills a puppy in your name.
Left MS Windows for Linux Mint and never looked back!
Vote for Bernie in 2016!
Fastest will probably be 'Servo', which is the new browser engine by Mozilla.
It's a parallel browser engine, it can render multiple parts at the same time:
https://www.phoronix.com/scan....
It's also written in Rust a language probably less prone to security issues than C++ (which all the other engines are written in).
A general article about Servo:
https://lwn.net/Articles/64796...
New things are always on the horizon
Try going to amazon.com, and just sit there on the front page doing nothing.
One entire core will peg out to 100%, the computer fans kick into turbo mode, and the browser process uses 2GB of memory.
Just sitting there doing nothing with JS enabled.
Nope. Don't see that at all.
Choose the right tool for the job.
:)
If you want normal UI DOM enables fast development, and isolates things nicely (compared to coding against a frame buffer).
But for rendering some things raw framebuffers (canvas) is and always will be the right tool...
Could also be you're using setTimeout instead of requestAnimationFrame...
But yes you're right, the DOM is slow... but you're not forced to use it when making pixel art...
he he, I think I once saw a js library to draw with 1x1px divs
Since Edge comes from the same company that makes the OS used to run those benchmarks, the fact the it did not win in all, or even most of, instances is a failure.
Um, No...
I'm sure that there is a trade-off where optimizing for specific content, which can have an effect on the loading time of other content. So, it would make sense for some browsers to have an edge depending on the test format.
Also, the benchmark includes checking for support for specific formats and functions, some more esoteric than others. For example, the HTML5 test includes a test for Ogg Theora. Mozilla and Opera were proponents of this format and support it, but it appears that it is being left in the dust for H.264 and VP9/10. Edge's lack of support for this format would hurt it in this benchmark, but in reality it's a moot point as very few web sites use this format.
"Edge gets 402 points on the HTML5Test.com test compared with 526 for Chrome and 467 for Firefox out of a possible 555 points. A lot of this is for code that very few sites use, such as Ogg Theora video, but some useful capabilities like Web notifications and WebRTC support for mic and camera access are still missing."
http://www.pcmag.com/article2/...
Take all benchmarks with a grain of salt...
I think the question is, which one allows you to install blockers and which one isn't spying and reporting everything you do back to the mothership?
-Unresolved symbol? Byte me!
The taste of irony.
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
With AdBlock active and preventing connections to tracker sites it's even better.
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
How about good adblockers for edge?
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
Why on earth would I care at all about "performance" in my web browser. Unless its 10x slower, seriously who cares?
What I care about are:
Obviously you can't get perfect in either, but I'll err on the side of coming closest to these marks.
Not sure if you are trolling or asking but since you brought it up I can say I've been using Waterfox as my primary browser for the last two years and it works well. I tend to run with upwards of 12 to 15 tabs at a time and it is quite responsive but does on occasion lose its connection to the network and require it be restarted to restore the connection. But since both Chrome and Edge no longer support NPAPI I suspect I'll be using it for the foreseeable future. At least until all my favorite web sites stop using features that require it.
"A person is smart. People are dumb, panicky dangerous animals and you know it." - K
Edge doesn't let me, or a script, print just one frame or iframe content or selection content. Making it useless
Right, because my #1 use case for the internet is to print web pages so a browser that can't do that is 100% useless. I'm glad I finally ran into the other person who uses the internet the same way I do.
I had to write this comment twice. I forgot that printing your post and writing my reply underneath it doesn't actually post anything.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
> Frankly I think browser programmers should not stop the spinning load icons until the javascript has stopped loading shit and the content of the page is actually visible. I'm so tired of "Oh look the browser thinks the page has finished loading...but it's wrong."
I understand your frustration. There is a good solution too, which is used on most projects I'm involved with. The technical reason for that is that very, very often the JavaScript manipulates things on the page, so it has to wait for the things to be there before it can start. If your JavaScript says "put this image in the footer", it HAS to wait until the footer is there, then run. So the page (the dom and elements mentioned in the HTML) has loaded when the spinner stops. The JavaScript is set to start immediately after.
Probably the best solution, what I use, is to make sure that the page is usable without JavaScript. When it says it's done, it's ready to use. Then JavaScript can add convenient but not critical features like autocomplete when it runs, if JavaScript is enabled.
Why? Do you think Microsoft's OS has some magic CPU cycles set apart for their browser?
Look, I'm sure it's nice to know how fast you can open a window or tab, run some piece of Javascript thingie, and close it, but that's got very little to do with how I browse, and a benchmark result of "23456" doesn't tell me anything.
Go fetch a Fark.com 168-hour page, open all the links in separate tabs, and tell me how much memory it's burning and whether the browser crashes. Once it's open, THEN go run your benchmark.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
So run it on Linux and OSX and see who wins. If the browser does not work there, then it really does not matter.
I prefer the "u" in honour as it seems to be missing these days.
I program ticketing solutions for a living. Ever printed a ticket to anything?
I develop ticket-purchasing web-sites. Ever printed a ticket to anything? Printing the hidden frame is helpful for the page to actually print what you want, and not the entire page.
And iframes are 19 years old, I believe you. Frames existed way before iframes. So I'm saying decades again.
And I often selected a paragraph or two and hit print, instead of printing the ten page article.
Using paper creates trees. Welcome to capitalism.
Dude, I've been doing it for 20 years. It's not hard to print 5" by 5" ticket in HTML. Welcome to liquid layouts from the 90's. HTML 0.9 is really good at that.
After the e-commerce of the purchasing, and the contact CRM of the mailings, and the scheduling software of the event conference, and the ad display network for the sponsors, and the box office sales systems and cash drawers, and the touch-screen kiosks, and the barcode scanners at the door, and the web-site selling the thing in the first place, and the private wifi network in the building that doesn't have a reliable one of its own.
After all that, the ticket's a report, but only if it can be printed properly.
Oh yeah, and the ticket has contact information, event information, and sponsor ads and a barcode right on it.
Yes everything output by a complicated system to a consumer is a report.
Just you, and the few people who also want to believe it's true.
I compared fresh installs on a bunch of systems, and Firefox was faster in all but a couple of cases. But in even those cases, comparing the nightly builds with the multi-process support to Pale Moon revealed Pale Moon to be much less responsive.
It's all down to two things: do you NEED the crappy old Firefox UI? If so, and you're loading up the stock Firefox with addons to get that effect, then Pale Moon might seem faster to you for now.
The other thing is the different ways the UIs lag. Some people prefer Pale Moon's lagging, others stock Firefox's. Some people even prefer Chrome's style of lagging, to the point where they don't think it lags at all.
So would you say that Edge is useless as a web browser, or that it doesn't work very well for your specific use case? Because those are 2 completely different things, and it sounds like you're trying to claim that it is useless as a web browser. Obviously it's not. What's more, Microsoft is aware of the bug you've found and has promised a fix.
Yeah, a bug. Not "things missing for no reason", like they made a design decision to remove that, but a bug. Keep in mind also that Edge is not the new IE, it is a new browser. They did not remove anything, they've only been adding features and fixing bugs.
But I'm sure you've already added your voice to the bug report so that Microsoft knows that it is affecting people, rather than just bitching about it on Slashdot. After all, you're a developer.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
Excellent, Smithers!
I'm a gonna go donate to noscript development right now. Kittens rule!
It's "removing" a feature not because it's IE, but because this is a feature that's been in every browser for ages. If you build an e-mail client today, and it doesn't support flagging messages, then you've removed a feature.
It's useless because if it's missing one vital feature, then it's missing many more. And since it's not my job to seek out bugs in other people's products, and it's not my job to solve them, then I have no interest in telling them. I work for my clients, and when this kind of thing happens, then I get to solve it for my clients, not for Microsoft. My clients pay me. It's that simple.
I'm not here bitching. I'm here explaining that benchmarks between full-fledged browsers that work, and mini crummy browsers that don't work, aren't worth shit.
Do what everyone else in the industry does and open another page with the printer friendly version.
It makes everyone's life easier, even Chrome and Firefox users.
Meh, they still haven't finished the Electrolysis project for multiprocess Firefox and they've been working on that since 2009, if it's in pre-alpha now it'll take a decade before it's usable.
Live today, because you never know what tomorrow brings
I doubt Apple will let you run competing browsers on their computer.
Chrome, Opera and Firefox are available on MacOSX and have been for a while. What are you talking about?
Write boring code, not shiny code!
I thought they knew how they want to do Electrolysis without breaking addons to much and how to fairly easily fix any broken addons.
But recently when they said they would move to supporting WebExtensions this could mean a delay of the transition.
It's the right idea in the long term of course.
I wonder what the impact will be on the number of users Firefox has if the advantage of addons ecosystem gets smaller.
New things are always on the horizon
Any user's decision for a browser is pretty much made already. There's no "browser war" to be had. That's a good thing: in the past it was like that because IE had terrible rendering issues, bad usability and common security issues. These days the overall browser landscape is less black and white, and for web developers it matters less which client the user is running.
Basically I see the choice of browser like this:
As a part-time web developer I'm happy that IE9 is almost dead; everything else is relatively inconsequential in comparison. From my personal perspective I'd like Edge to be just slightly more competitive. Firefox is getting worse all the time (bad performance, terrible reliability, increasing bloat, breaking of old features like Firefox Sync). Google is too spooky for me to switch to Chrome. Edge would be very interesting if it was just a little bit better.
there is no winner in the benchmark test, so firefox is the clear winner.
it's open source and it is not being pushed by a for-profit company.
On a long enough timeline, the survival rate for everyone drops to zero.
Servo is still a year away, and Blink/Chrome has had parallel rendering for a long time now. Chrome threads everything. Network, layout elements, compositing, JS, even GPU offloaded tasks.
In other words, Mozilla are just catching up.
Firefox does okay in benchmarks, but it feels slow compared to Chrome and even to some extent Edge. The problem is this lack of threading. The page may render about as fast as Chrome but the UI is frozen while it does. It's particularly noticeable when you use the back button and a load of images and JS are reprocessed. JPEG decoding happens in the same thread as the UI and every other tab.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
I care far more about browser stability than browser speed.
Your definition of "doesn't work" is highly skewed. When someone sends you an email and says that the large system you've implemented "doesn't work", what's your response? I know what my response is to that non-bug-report. I ask them exactly what they're trying to do and what happens, because I know that the system as a whole works and they've just found a bug in some part of it. But, here you are, claiming that Edge "doesn't work". You're not filing bug reports to make sure the problem gets fixed on their end (thereby increasing the usability of your software on their product, in line with other products), you're just bitching.
It's useless because if it's missing one vital feature
This shows your skewed perspective. If that feature was as vital as you think it is then it would have been caught pre-release. If it's vital to you, then do something about it to make sure they fix it.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
Again, it's not my job to fix their product. They aren't paying me.
You are correct that my clients enjoy the "doesn't work" report. There's a reason that they all do it. It "doesn't work" for their business. It's not only true of bugs. It's also true of things being the wrong colour, or a missing feature. If it doesn't work for their business, then it simply doesn't work.
Asking them for more details isn't a part of their report. It's a part of your/my solution. In my world, if a client says it doesn't work, and I don't fix it, they stop using my product, stop paying me, and go elsewhere. The vast majority of "doesn't work" issues are deal-breakers in my industry.
I'm not interested in making sure that microsoft fixes it. I don't benefit from that fix.
You've missed the entire point here. The original post is about benchmarking edge versus other browsers. My point was that you can't benchmark a partial browser against a complete browser.
Stay in context.
Again, it's not my job to fix their product. They aren't paying me.
Filing a bug report and adding your voice to the list of people wanting it fixed is in no way, shape, or form you fixing their product. They have programmers to fix the thing themselves. The only thing the bug report accomplishes is helping to make sure that the problem gets fixed quicker, if that's of any concern to you.
I'm not interested in making sure that microsoft fixes it. I don't benefit from that fix.
You don't benefit from your product working the way it was designed in the default browser of Windows 10? Well, OK.
My point was that you can't benchmark a partial browser against a complete browser.
You sure as hell can when the benchmarks measure things like Javascript and DOM performance and HTML 5 accuracy instead of the ability to print an iframe. It's just a stupid claim to state that the browser is useless because it has a bug that affects your particular application or use case, which you have incorrectly categorized as a missing feature. It's like if I tried to claim that since I use Opera, and it has mouse gestures, that any other browser that does not have mouse gestures is "useless". It's a stupid claim to make, because there are clearly many people out there using browsers that simply do not use them the same way I do.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
There's a difference between mouse gestures, which are a rare feature, relatively new, and are a user-selected, per-user feature, and printing, which is a cross-user, fundamentally long-standing feature, scripted away from the user. The former is how it's used, the latter is what it does.
You say they "have programmers to fix the thing themselves". Well, they also have testers to find the problems themselves and analysts to prioritize the problem themselves. You can pray if you want to; I don't.
And no, there's no benefit to me, professionally, from my product suddenly working in edge when before it did not. In fact, there are two huge on-going benefits every day that it remains broken -- welcome to business: profit and service.
Usability is a theshold scenario. Countless tiny things are tiny until they add up to something that crosses that threshold, at which point the entire item becomes useless. For me, mine, and those around me, a browser that can't print is that threshold -- making it useless.
There's a difference between mouse gestures, which are a rare feature, relatively new
When Opera released a version supporting mouse gestures, they were competing with IE 5. There are high school kids younger than mouse gestures.
The former is how it's used, the latter is what it does.
Way to completely miss the point.
Well, they also have testers to find the problems themselves and analysts to prioritize the problem themselves.
That's correct. And guess why they didn't find this particular bug before release, or why they decided to release it anyway before fixing the bug (hint: it's the same reason!). Go ahead, guess.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black