[webkit-dev] setTimeout as browser speed throttle

[Maciej Stachowiak]

On Sep 30, 2008, at 3:30 PM, Peter Kasting wrote:

> On Tue, Sep 30, 2008 at 3:08 PM, Maciej Stachowiak <mjs at apple.com>  
> wrote:
> That seems like incorrect reasoning to me. unclamped setTimeout(0)  
> does not break processing of user events in a single-process browser  
> (I tested). But it will equally drain your laptop battery and  
> produce a great deal of heat and noise with single-process and  
> multiprocess architectures.
>
> Right, which is one reason we don't run fully-unclamped -- we clamp  
> to 1 ms, and are trying to get anyone other than hyatt to discuss  
> the ramifications of clamping to 3 or 4 ms.  I think this proposal  
> solves most of the objections that have been raised, but people  
> still seem more interested in discussing the problems encountered  
> back when WebKit had fully-unclamped timers, which is a different  
> and more risky case.

I don't think there is a huge difference between 1ms and no clamp at  
all. 3ms or 4ms would be significantly different (probably). But it  
also seems less worth it to make such a change if a high resolution  
unclamped API is added as well.

> I certainly do support the proposals for an object-based, higher- 
> resolution timer alongside.  But the responsiveness/performance gain  
> from lowering the timer clamp on setTimeout() is noticeable.  It's a  
> real win, which you can see today, on a number of existing sites.

Can you cite some of the existing sites that would benefit? That would  
help others confirm the benefit and also estimate likelihood of said  
sites adopting a new better API for greater benefit.

Regards,
Maciej

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.webkit.org/pipermail/webkit-dev/attachments/20080930/b1256d71/attachment.html