[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnash-dev] Re: performance: 1000 wakeups/sec when idle!

From: Yuriy Kaminskiy
Subject: [Gnash-dev] Re: performance: 1000 wakeups/sec when idle!
Date: Fri, 05 Jun 2009 17:48:54 +0400
User-agent: Icedove (X11/20090105)

strk wrote:
> On Fri, Jun 05, 2009 at 03:32:54PM +0400, Yuriy Kaminskiy wrote:
>> This is not right at all. Unless you have running timer (and no timer
>> run in paused mode), you should not set timeout in poll() at all
>> (something like poll(nfd, fds, -1)); or, at least, set it way higher.
> Any chance to also find the spot in gnash calling poll() ?

Think, it traces to
gui/gtk.cpp:GtkGui::setInterval(unsigned int interval)
and similar methods for other frontends: gnash always run that periodic
low-delay timer, whether it needs it or not.
And that causes those unnecessary wakeups.

Plus, there may be more problems with /unpaused-by-gnash/ movies: can we
avoid running that timer for static pictures? Say, this one:
{looking in swfmill - not easy, it seems busylooping inside flash code};
or on paused-by-flash-||-button youtube movie? Well, I don't know gnash
code/flash internals to dig deeper.

> Also, which movie was gnash running during your strace calls ?

Random thing found on LAN ;-) [after some googling] Seems, this one,
Well, judging by above, that does not matter, gnash would behave same on

reply via email to

[Prev in Thread] Current Thread [Next in Thread]