Comments on: Backbase tests browser JavaScript and Render performance including IE 8

By: JepCastelein

JepCastelein — Fri, 07 Mar 2008 21:06:57 +0000

OndraM, the results of the first test that you link to are actually quite similar. Our figures also show that Firefox nightly is faster than Safari 3.0.4b, and that Safari 3.0.4b is close to Firefox 3 beta 3 (our test shows Safari slightly faster, the other test Firefox). We spent a lot of time optimizing our code for IE, so maybe that’s why IE looks faster in our test compared to Firefox 2. Opera is an anomaly because we tested event propagation extensively, and Opera is apparently not very fast in that.
Of course, the results of any benchmark depend on the exact operations you include, so there will always be variation. Nevertheless, it’s clear that both Firefox 3.x and Safari 3.x are really fast, and getting faster in the nightly builds. IE is still pretty slow, but getting better. Not sure about the Wired News test results: they don’t seem very scientific.
@ Alexei: we’ve used IE7 mode, but we’ll do some further testing in the other modes.

By: alexeiwhite

alexeiwhite — Fri, 07 Mar 2008 17:15:29 +0000

The thing about the IE8 tests is it really seems to matter what mode you’re running it in.. whether its the new IE8 standards mode, or IE7 mode and so-on. My experience (at least with the beta build) is that the rendering performance in ie8 standards is Abysmal. I’m curious what other people have experienced.

By: OndraM

OndraM — Fri, 07 Mar 2008 15:31:46 +0000

Hi Jep, one was recently here on Ajaxian:
http://ajaxian.com/archives/firefox-3-performance-numbers

I guess it depends quite a lot on chosen methodology. How else can one explain this bizzare thing ;)
http://blog.wired.com/monkeybites/2007/06/wired_news_benc.html

I went through the methodology of the Backbase and detailed results – of course – after posting my previous message :( – and it seem very thorough, however.

By: JepCastelein

JepCastelein — Thu, 06 Mar 2008 21:10:39 +0000

Hi OndraM, can you provide some links and more detail? Please note that 3 of the browsers tested have been released in the last week, so I doubt there will be many tests around. Thanks, Jep

By: OndraM

OndraM — Thu, 06 Mar 2008 16:36:09 +0000

interesting results – however they do not correspond to most speed tests I have came across before….

By: JepCastelein

JepCastelein — Thu, 06 Mar 2008 16:31:26 +0000

Hi Anonymous, I encourage you to read the article at http://bdn.backbase.com/blog/sjoerd/performance-is-everything: it gives the reason for Opera seeming slow: “the main reason for this is the test for the event propagation being 5x slower in Opera than Firefox”.

By: Anonymous

Anonymous — Thu, 06 Mar 2008 15:53:02 +0000

Even if the y-axis is time, the graph still doesn’t make any sense. That would mean Opera is the slowest of the bunch, while in reality it has one of the fastes js engines available.

By: JeromeLapointe

JeromeLapointe — Thu, 06 Mar 2008 15:36:45 +0000

So the shorter the bar the better right?

By: leveille

leveille — Thu, 06 Mar 2008 15:36:29 +0000

Anonymous, I think you need to take another look at the graph. Safari is actually represented very well.

FireFox 2 has been a pig for me as of late. I am also eagerly awaiting the public FF3 release.

By: pwnedd

pwnedd — Thu, 06 Mar 2008 14:36:47 +0000

A label on the y-axis would be helpful. From the original article, the y-axis represents “the time it takes to execute the different engine performance tests.” Units wouldn’t hurt either.

By: Anonymous

Anonymous — Thu, 06 Mar 2008 14:32:19 +0000

Those graphs don’t seem to make any sense. For example, Safari’s js engine is no where near as slow as it’s displayed there.