Re: Speed suggestions (previously about laptop performace)
- From: Michael Meeks <michael ximian com>
- To: Soeren Sandmann <sandmann daimi au dk>
- Cc: Telsa Gwynne <hobbit aloss ukuu org uk>, gnome-list gnome org, gnome-hackers gnome org
- Subject: Re: Speed suggestions (previously about laptop performace)
- Date: 05 Aug 2002 13:48:09 +0100
Hi Soeren,
On Mon, 2002-08-05 at 10:55, Soeren Sandmann wrote:
> One problem with both methods is that the top of the stack is not a
> normal frame when you are inside a signal handler....
Sure; I just use memprof's existing bin the top bit of the stack code
which skips that.
> which means that leaf procedures are never accounted for. This could
> account for the lying.
I'm talking about really, really acute lying. For example:
gdk_flush is no. 4 on the list, apparently most often called from
gdk_image_get_colormap:
GdkColormap *
gdk_image_get_colormap (GdkImage *image)
{
g_return_val_if_fail (GDK_IS_IMAGE (image), NULL);
return image->colormap;
}
That's the sort of lying I'm talking about; not some slightly skewed
numbers. Amazingly, some of the data is still useful despite this.
Regards,
Michael.
--
mmeeks gnu org <><, Pseudo Engineer, itinerant idiot
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]