Slow konqe on a PXA250

[Daniel Stenberg]

On Wed, 20 Nov 2002, Mayuresh Kathe wrote:

> > For evaluation purposes, I'm currently running Linux on an Intel XScale
> > PXA250 Processor (on one of them Intel development boards)

> Sounds good, where and how much did you buy that for?

We're currently borrowing this one (or in fact, we borrow two at the moment
as we thought the second set would have the 100MHz memory fixed, which it
didn't). But I think it is priced at around 5000 US$ and we'd buy it from an
Intel sales place (the board itself is produced by Accelent).

> > I have konqueror-embedded-snapshot-20020311 built and running on top of
> > the Qt/E 2.3.2. (I have not done any particular optmizations or
> > configurations, merely got them working using the smallest amount of
> > effort possible.) I run the browser full-screen.
>
> Optimisations are required for smaller binary or lesser resource usage.

Right, and I'll come to that. I just don't think that the problems I see here
is related to that. I'll certainly optmize later on.

> Hope you have atleast used the recommended JPEG and PNG libs.

Which ones are the recommended? I use the one bundled with Qt.

> > The system has 64MB ram and 10Mbit ethernet connection.
>
> Excellent hardware for Konq/e :-)

This is only the devboard's setup. We're gonna go for 32MB ram in the final
product. Initial tests prove that we can do everything we want within that.

> > I enter the URL for that HTML file in konqueror and press return. I can
> > see in my ethereal trace that the fetch of the HTML file flies by really
> > fast. (The http server is connected on the same hub as the devboard.)
> >
> > Then it can take up to almost 3 seconds of no network traffic at all
> > before konqueror decides to get the first image. It then gets the second
> > and third images within a quite reasonable time (fractions of a second),
> > just to again make a pause before the 4th image is retrieved. This
> > behavior is easy to repeat and seem to be roughly the same whenever I
> > test this.
>
> Thats funny :-(
> No network traffic for 3 seconds after issuing a fetch is an impossible
> state (mind you, I am not being derisive of your observations).

You're only confirming my own reactions. I couldn't believe my eyes the first
times and I'm still struggling to find explanations and reasons behind this!

> The only thing that comes to my mind is that there must be network traffic,
> just that Konq/e might be taking 3 seconds to get all the images to its
> cache before displaying them.

My ethereal trace shows that this is not the case. The network is not used
for anything else. There are only the client and the server connected on it!

> Secondly, Konq/e (AFAIK) doesn't do progressive display of content like
> rest of the web browsers, it first gets everything and then displays it all
> at one go (am I right Simon and Paul?)

It clearly shows one image at a time, and for web pages that contain both
text and images I can very easily detect that it first displays the text and
then re-writes the display when the image arrives. So a certain degree of
progressiveness is there all right.

Hm, does Konq/e create new threads often? I found a post over at the
armlinux-kernel list that identified a problem in the Linux process creation
code for the PXA...

-- 
      Daniel Stenberg - http://daniel.haxx.se - +46-705-44 31 77
   ech`echo xiun|tr nu oc|sed 'sx\([sx]\)\([xoi]\)xo un\2\1 is xg'`ol