Re: Gnome first-time login profiling report.
- From: Michael Meeks <michael ximian com>
- To: Malcolm Tredinnick <malcolm commsecure com au>
- Cc: desktop-devel-list gnome org
- Subject: Re: Gnome first-time login profiling report.
- Date: 27 Aug 2002 12:14:51 +0100
Hi Malcolm,
On Tue, 2002-08-27 at 00:44, Malcolm Tredinnick wrote:
> To me, at least, this is interesting stuff. I have one question, though:
> when you say "first-time" login, do you mean that the user in question
> has never logged into GNOME with that home directory, etc, before (so
> .gnome2 and so forth are being created).
Nope - this is first time after a reboot. This is primarily so things
such as OS caches are not primed, the disk heads have to move to get
data etc. It's also ( of course ) the time that most people will
experience as the worst bound of login time.
> Is it possible to put up a tarball of the patches you made (or just a
> description) where you added debugging output? I would like to run
> similar tests on some Linux systems to see if the ratios are about the
> same.
I think for a Linux system - they are fairly similar, yes. Some things
are significantly worse for Solaris systems such as X font queries - but
in the main, optimising on the basis of the Solaris trace will also
yield siginificant gains on Linux.
Regards,
Michael.
--
mmeeks gnu org <><, Pseudo Engineer, itinerant idiot
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]