Chrome now outperforms Safari in Javascript benchmarking

[*LTD*]

Performed 3 tests using Sunspider.

Chrome (5.0.307.7 beta)

Test 1: 409
Test 2: 405
Test 3: 397

Safari (4.0.4)

Test 1: 496
Test 2: 495
Test 3: 493

In terms of personal experience, Chrome feels a lot faster now.

http://www2.webkit.org/perf/sunspider-0.9/sunspider.html

 

[clevin]

and opera 10.5 beta 1 outperform them both by 300-500%.

 

[SHADO]

Let's face it: most browsers are faster than Safari right now (Chrome, Opera, Firefox to a certain extent). Apple can do a lot better in terms of optimizing Safari's speed that they have yet to do. And they better do it fast or Safari will just keep losing to the rest of the browser field.

 

[Melrose]

I very much prefer Safari for everything other than it's initial boot up. It takes like :30 seconds for the browser to load, and Chrome is a snap, literally.

I really wish Safari would just start up faster. The extra few milliseconds doesn't make that big of a difference to me. That, and scrolling Chrome is wacked up.

 

[angelwatt]

Very few people notice the speed difference. If they did, do you think IE would still be the most used browser (notice I didn't say popular)? For me, it's about features, not speed. They all have good enough speed in terms of JavaScript. I have most browsers installed for web development testing and find nice things about each one. The addons keep me with Firefox though for the majority of my web browsing. The latest Chrome is more usable now, especially since it has an actual bookmark manager finally and can use extensions (not on par with Firefox addons yet).

 

[str1f3]

I very much prefer Safari for everything other than it's initial boot up. It takes like :30 seconds for the browser to load, and Chrome is a snap, literally.

I really wish Safari would just start up faster. The extra few milliseconds doesn't make that big of a difference to me. That, and scrolling Chrome is wacked up.

If Safari is taking 30 seconds to load, then that is your fault. I suggest you get rid of all those plugins that you're using. Safari takes one bounce on my dock to load and I only use Click2Flash. I suspect that this is part of the reason why LTD got the results he did (compared to a fresh install of Chrome) when every other benchmark shows them neck and neck.

BTW, I speak from experience. I used to have Glims, Inquisitor, and a whole bunch of SIMBL plugins giving me 30 second loading times. Browser speeds are not only dependant on Javascript performance but the user as well.

 

[clevin]

I suspect that this is part of the reason why LTD got the results he did (compared to a fresh install of Chrome) when every other benchmark shows them neck and neck.

i suggest you check title again, the op is a javascript benchmark, fresh install or not makes absolutely no difference.

 

[str1f3]

i suggest you check title again, the op is a javascript benchmark, fresh install or not makes absolutely no difference.

Yeah, I fully understand the topic and, yes, many plugins do affect the Javascript performance especially with Safari which is really not designed to handle plugins. And yes, compared to fresh install of a browser with no plugins, does make a difference. Just by the mere fact that many tests from established tech sites show scripting performance basically equal between the two, says that there are other issues at hand that is affecting performance.

 

[xUKHCx]

Of course they both have rendering issues on that specific javascript benchmark site which I would say is more important.




Safari, Chrome, Opera

 

[plinden]

Of course they both have rendering issues on that specific javascript benchmark site which I would say is more important.

View attachment 214689


Safari, Chrome, Opera

Are you sure it's Opera that's rendering correctly? Considering Sunspider is hosted on the WebKit site, you'd expect both Safari and Chrome to render correctly. WebKit itself displays the same as Safari and Chrome.

I've benchmarked JavaScript performance in the past, and how long your computer has been running is a big factor in measurements. I always do several runs and discard the first one, and reboot before testing each browser.

 

[iWoz]

Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 3_1_2 like Mac OS X; en-us) AppleWebKit/528.18 (KHTML, like Gecko) Version/4.0 Mobile/7D11 Safari/528.16)

Safari and MobileMe need improving!!

 

[clevin]

Yeah, I fully understand the topic and, yes, many plugins do affect the Javascript performance especially with Safari which is really not designed to handle plugins. And yes, compared to fresh install of a browser with no plugins, does make a difference. Just by the mere fact that many tests from established tech sites show scripting performance basically equal between the two, says that there are other issues at hand that is affecting performance.

js development changes fast, and most tech site i saw says chrome js speed is better. You said the sites you read showing these two equal, do share some links.

if you have this much problem with op's result, why don't you test it yourself and post them? Its not that hard to do.

Are you sure it's Opera that's rendering correctly? Considering Sunspider is hosted on the WebKit site, you'd expect both Safari and Chrome to render correctly. WebKit itself displays the same as Safari and Chrome.

I've benchmarked JavaScript performance in the past, and how long your computer has been running is a big factor in measurements. I always do several runs and discard the first one, and reboot before testing each browser.

it doesn't matter how long the computer is running, it affects both, so the relative difference is still valid.

im curious about sunspider's design now

 

[netnothing]

im curious about sunspider's design now

The reason for the differences in the sun is because it's not a graphic, it's a Unicode character that will be rendered differently between browsers and OSes.

http://unicode.org/charts/nameslist/n_2600.html

<h2><span id="logo">☀</span>SunSpider JavaScript Benchmark</h2>

Also differences in how the browsers render the CSS on the page will result in the different placement on the page.

-Kevin

 

[clevin]

The reason for the differences in the sun is because it's not a graphic, it's a Unicode character that will be rendered differently between browsers and OSes.

http://unicode.org/charts/nameslist/n_2600.html

<h2><span id="logo">☀</span>SunSpider JavaScript Benchmark</h2>

Also differences in how the browsers render the CSS on the page will result in the different placement on the page.

-Kevin

so whats the reason for Opera's difference rendering on windows 7 vs mac osx?

 

[netnothing]

so whats the reason for Opera's difference rendering on windows 7 vs mac osx?

The OS is the difference. So many things in CSS are different between Windows and OS X. When it comes to fonts and stuff like line height, the two OSes will display stuff differently, even in the same browser.

-Kevin

 

[clevin]

The OS is the difference. So many things in CSS are different between Windows and OS X. When it comes to fonts and stuff like line height, the two OSes will display stuff differently, even in the same browser.

-Kevin

well, then why does firefox (or chrome) display it the same on all platforms?

look carefully, opera on win7 not only has thicker lines, but actually MORE lines.

 

[Hellhammer]

You need a second watch to notice the difference between today's browsers...

 

[netnothing]

well, then why does firefox (or chrome) display it the same on all platforms?

look carefully, opera on win7 not only has thicker lines, but actually MORE lines.

No idea. I don't use Opera (or even care that it exists). Why we can't just have all browsers use the same rendering engine I'll never understand.

If you are really concerned you should send a request off to Opera so they can tell you how they do it right an the rest of the world is wrong.

-Kevin

 

[Aronnax]

says that there are other issues at hand that is affecting performance.

And sometimes only benchmarks

For example: http://trac.webkit.org/browser/trunk/WebCore/dom/Document.cpp#L1828

// Make sure both the initial layout and reflow happen after the onload
// fires. This will improve onload scores, and other browsers do it.
// If they wanna cheat, we can too. -dwh

---------

By the way,
some know the lie and use it anyhow for advertising
For example: http://www.apple.com/safari/whats-new.html#performance

// Just bail out. Before or during the onload we were shifted to another page.
// The old i-Bench suite does this. When this happens don't bother painting or laying out.

 

[clevin]

No idea. I don't use Opera (or even care that it exists). Why we can't just have all browsers use the same rendering engine I'll never understand.

If you are really concerned you should send a request off to Opera so they can tell you how they do it right an the rest of the world is wrong.

-Kevin

im not concerned, im just curious.

why we can't just have all browsers use the same rendering engine? well, last time we tried, we had IE. we all know what happened since.

 

[netnothing]

why we can't just have all browsers use the same rendering engine? well, last time we tried, we had IE. we all know what happened since.

Sadly as a web developer I do remember.....and am reminded everyday when I have to sit here and code tweaks for all versions if IE. Talk about holding the web back!

-Kevin