Fernando Correia’s Weblog

September 2, 2008

Chrome is Fast: JavaScript benchmark

Filed under: Software Development — Fernando Correia @ 5:29 pm
Tags: ,

I ran the V8 JavaScript virtual machine’s benchmark suite on Internet Explorer, Firefox and Chrome. Guess who comes first…

The old JavaScript engines are not even in the same league. For a competitor I think we must wait for Tamarin TraceMonkey.

Edits:

Reader PA contributed with another benchmark that finds that Lua is faster. Good, let’s all switch to Lua then! =D

Another independent benchmark: Google Chrome browser mini-review, benchmark (still fast).

Something more scientific-looking here: Google Chrome Benchmarks (still rocks).

Chip has a nice review and uses Mozilla’s own benchmark (still flies):

Alex pointed out that Celtic Kane has an interesting benchmark that deserves a good look. It shows very different results.

(Sorry, but Opera is not my style…)

Also John Resig did a great job on several benchmarks across several browsers:

Brendan Eich speaks for the TraceMonkey team with some in-depth analysis.

Matt Hackett also published a benchmark.

Faster JavaScript and faster browser engines mean richer browser applications without plugins like Flash and Silverlight.

Advertisements

19 Comments »

  1. […] Edit: See the benchmark. […]

    Pingback by Google Chrome is Out « Fernando Correia’s Weblog — September 2, 2008 @ 5:38 pm

  2. In fairness, it isn’t absurd to imagine that V8’s benchmark suite might favour V8. Results on other benchmarks are still very impressive, though, just not AS impressive.

    Comment by Robert Synnott — September 2, 2008 @ 5:46 pm

  3. Robert: I agree, the code will end up being adjusted to whatever rule is being used to measure it. I know my approach is not scientific, but I think these results still mean something. And they did pick old algorithms written by other people, so it’s probably not too biased.

    Comment by fernandoacorreia — September 2, 2008 @ 5:55 pm

  4. Another benchmark, comparing various JS VM:

    http://article.gmane.org/gmane.comp.lang.lua.general/50131

    Comment by PA — September 2, 2008 @ 6:02 pm

  5. You should also make sure that you’re benchmarking against Mozilla’s beta browser since the Chrome browser is also beta. The next alpha release from Mozilla will include TraceMonkey which contains similar technology. Moz folks are still digging into the tests and will have more to say on the topic, I’m sure.

    Comment by Christopher Blizzard — September 2, 2008 @ 6:23 pm

  6. Christopher: My “benchmark” run is not scientific at all. And I’m using Firefox 3.0.1. And Google products are always beta. Beta is their way of saying “use it”.

    Comment by fernandoacorreia — September 2, 2008 @ 6:30 pm

  7. I don’t think you should wait for Tamarin (JS4 being dead, Tamarin is pointless). You should test with TraceMonkey:
    http://ejohn.org/blog/tracemonkey/

    Comment by Erwan — September 2, 2008 @ 6:32 pm

  8. Erwan: You’re right, of course. Thanks.

    Comment by fernandoacorreia — September 2, 2008 @ 6:37 pm

  9. I’ve run a comparison against JavaScriptCore on OS X. V8 comes out ~4.6x faster than the latest JavaScriptCore build when running the benchmark suite:

    http://www.tangerinesmash.com/writings/2008/sep/02/v8-and-javascriptcore-compared/

    Comment by Andrew — September 2, 2008 @ 7:28 pm

  10. Not for me. I wrote a Reversi Game in JS and FF 3 is still much faster, about double speed. It really seems to depend on the code.

    Comment by Searle — September 2, 2008 @ 8:02 pm

  11. […] The V8 Javascript engine in Chrome is really fast – thanks to just-in-time compilation and other optimizations. This is important, because it […]

    Pingback by Tim Anderson’s ITWriting - Tech writing blog » Chrome: the developer angle — September 2, 2008 @ 9:08 pm

  12. And where is Opera???

    If you run that test on Opera you will see that is faster than the others browsers you mentioned: FireFox, Safari and IE, of course, Chrome is still the fastest. Why you didn’t mentioned Opera? The second fastest browser… and the fastest browser that have currently a stable release. Aaaa… you are a FF fan… and you wouldn’t spell that out!

    Comment by alex — September 3, 2008 @ 8:48 am

  13. Alex: Hi! I guess Opera is where it belongs. With its users. At least one of the links in my page refers to a benchmark that includes Opera. And it loses. Badly. About being a Firefox fan, I don’t think it needs to be said, I take it for granted. What would I use instead? Opera? Well, when Chrome has one or two plugins I want, I’m leaving Firefox behind.

    Comment by fernandoacorreia — September 3, 2008 @ 9:36 am

  14. Well… taking it that way, I think that every browser is with its users, except IE. I was referring to where is Opera on your test graphs?

    There are a lot of tests out there on the web that checks different performances of the browsers, so almost every test has a different top 3. You can take this JS test for example celtickane[dot]com/webdesign/jsspeed.php in this case Opera is faster even then Chrome.

    But, what I wanted to say, if you take a particular test, here you have the v8 benchmark from the Google Code website, and you want to present an objective perspective to your readers, you should include all the major web browsers, and Opera is one of them, despite that your favorite surfing board is losing or not a place on the podium.

    Comment by alex — September 3, 2008 @ 10:33 am

  15. Hi Alex! Thanks for your feedback. Let me try to explain. I’m not trying to present any objective perspective here. I’m publishing a single benchmark on my machine with the browsers I happen to use. As about what I “should” do, that’s for me to decide. I’m pointing to other sites that have better benchmarks than mine. As I see, Chrome is fast, and I think faster JavaScript on several browsers will change the game. I’m sure soon we’ll see a lot of objective data on all that. Cheers!

    Comment by fernandoacorreia — September 3, 2008 @ 11:32 am

  16. Is the compilation time included in the benchmarks?

    Comment by Fred — September 6, 2008 @ 10:41 am

  17. Sorry, Fred, I don’t really know. I ran it twice on each browser and used the last result. My test is too simple anyway; see the linked articles for something more elaborate.

    Comment by fernandoacorreia — September 6, 2008 @ 10:47 am

  18. Hi! I was surfing and found your blog post… nice! I love your blog. :) Cheers! Sandra. R.

    Comment by sandrar — September 10, 2009 @ 11:27 am

  19. A little off the subject maybe, but a plea for people to consider the ethics of purchasing items like designer clothes. Please try and think about, for example, the materials your item is made from, the conditions of the factories where they’re manufactured and the green credentials of retailers. Oh, and endeavour to recycle instead of throwing away. Thanks!!!!

    Comment by Vanessa Bruno — April 27, 2010 @ 4:39 am


RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at WordPress.com.

%d bloggers like this: