4-way HTML5 speed test: Firefox 3.7 faster than Internet Explorer 9 (video) | Hacker News Reader

4-way HTML5 speed test: Firefox 3.7 faster than Internet Explorer 9 (video)

downloadsquad.com

23 pointsxaverius7 comments

This is getting silly. It's fantastic that we finally got canvas in Explorer, but people are putting too much weight to one demo designed to highlight one particularly well performing feature (HW accelerated image draw on canvas #).

This has as much significance as Chrome getting ridiculously better results in V8 benchmark [1] compared to all other browsers.

Please go to CanvasDemos [2] or ChromeExperiments [3] and try some third party canvas demos to see how well IE9 fares across wider variety of canvas use cases.

Spoiler alert: it's nowhere near so clear cut performance champion as it may seem from fish and asteroids demos.

----

Edit: (#) in practical terms, the biggest benefit would go to use cases like the upcoming Aves JS game engine [4] and FreeCiv JS port [5].

----

[1] http://v8.googlecode.com/svn/data/benchmarks/current/run.htm...

[2] http://www.canvasdemos.com/

[3] http://www.chromeexperiments.com/

[4] http://www.dextrose.com/en/projects/aves-engine

[5] http://www.freeciv.net/

I don't think anyone notices Javascript execution performance at this point, but without the hardware acceleration capabilities of Firefox and IE9, HTML5 canvas has no chance of replacing any use-cases of flash.

Chrome is not the fastest.

Sunspider benchmark (Win XP, dual core machine):

Opera (v10.53): 299.8ms +/- 1.0%

Safari (v5): 314.2ms +/- 1.7%

Chrome (v5.0.375.55): 326.6ms +/- 7.9%

Firefox (v3.6.3): 716.4ms +/- 1.7%

You didn't read carefully - not Sunspider, V8 benchmark:

  Firefox 3.6.4       494
  IE9                1200
  Safari 5           2644
  Opera 10.54        3393
  Chrome 6.0.437.3   5148

Saying that Chrome is 10x faster than Firefox on the basis of V8 benchmark would be as misleading as saying IE9 is "orders of magnitude" faster than Chrome on the basis of fishtank demo.

Though it certainly makes great headlines.

What a terrible test..

When the browsers tried to get to 1000 you could see them fluctuating as the os gave and took their resources (even slightly noticeable at 100 (mainly in opera)). They needed to be done one at a time at that point.

On the same computer... at the same time? That seems a little unfair if one browser is choking up the disk access (as Firefox seems to do on me every few seconds).

Disk access? What? I/O has stunningly little to do with this.

Reply on news.ycombinator.com