Internet Explorer Canvas Tag Massively Faster Than Chrome(nerds-central.blogspot.com)
nerds-central.blogspot.com
Internet Explorer Canvas Tag Massively Faster Than Chrome
http://nerds-central.blogspot.com/2011/10/ie-kicks-back-and-wipes-floor-with.html
2 comments
Following up, the site uses "setTimeout(function, 0)" which is akin to an unencumbered while() loop. Great if you want raw power but pointless if you cannot use it (in this case, to display unique data).
Perhaps Chromium 16 has tied sites that incorrectly use setTimeout() for animations to this newer method: http://dev.chromium.org/developers/design-documents/requesta...
Perhaps Chromium 16 has tied sites that incorrectly use setTimeout() for animations to this newer method: http://dev.chromium.org/developers/design-documents/requesta...
I've found IE to be faster too.
Usually when developing web apps I get things working in Firefox first then test in the other browsers.
Lately I've been doing some work with Canvas and found IE to be the fastest for drawing large (many points) polygons(* ). I never thought I'd say this but it's been a lot more pleasant to develop in IE, at least during the proof-of-concept stage where I haven't done much optimizing or caching.
(* ) I haven't benchmarked whether it's the canvas drawing or javascript engine (or both) that's faster, just that overall I'm getting 30+% better performance from IE
Usually when developing web apps I get things working in Firefox first then test in the other browsers.
Lately I've been doing some work with Canvas and found IE to be the fastest for drawing large (many points) polygons(* ). I never thought I'd say this but it's been a lot more pleasant to develop in IE, at least during the proof-of-concept stage where I haven't done much optimizing or caching.
(* ) I haven't benchmarked whether it's the canvas drawing or javascript engine (or both) that's faster, just that overall I'm getting 30+% better performance from IE
In other words, the code at http://www.nerds-central.com/Linear-Scan-Test.html looks like this code is using the antiquated setTimeout() method of determining frame rate.
A better way to do it, at least for chrome and firefox users, is to use the newer requestAnimationFrame() function: http://paulirish.com/2011/requestanimationframe-for-smart-an... This allows for the browser to notify the script when its ready to display a new frame, resulting in more efficient animation execution.
Using FPS for benchmarking is good, but suffers from the same shortcomings as glxgears.