innerHTML performance in multimonitor system

[ppRaistlin]

Hi,
I understand that it s a very rare question that I m ready to ask and
that only few people can actually test but a theoric explanation and
solution will be really grateful.

Here is my problem :

I have a computer with 3 video cards and 6 monitors plugged (for
business purpose). The resolution for each monitor is 1280x1024
On each of this monitor I have a IE browser (full screen) showing the
same html local page.
This HTML is very simple containing a 2 <img> (640x480 for one and the
other is very small) , 15 div nested on 3 three levels only containing
static text.
The content of the first of these div (this one is not nested) is
refreshed every 1 s (by a settimeout) with innerHTML
document.getElementById("MyDiv").innerHTML= "Hello"
The HTML contains no javascript except the settimeout (started onload)
and is linked to a CSS. I don t know if it s important but the size of
every div is in % and the fontsize is em

The problem is that the process time used by each browser go to 11 %
and stay at this value, so a permanent total of 66 % for a simple
recurrent innerHTML (stay at this amount as long as the browsers are
running) on 6 browsers. I find this value very high and unlucky for me
non compatible for the use i want to made with my application.

I have tried other solutions like the well known "replaceHTML" or DOM
method without any success. The only way I have found to take down the
process time to 0 %!!! for every iexplorer.exe and keep my recurrent
innerHTML refresh, is to open 6 new browsers containing only one div
and this timer (and remove the div and the timer from the first 6
browser). I can conclude that the problem is caused by using a
recurrent innerHTML on my simple html page.

Do you think that's normal ? Have you any solutions or ideas about
that ?

My configuration is WinXP 2 Go, Athlon 64 X2 3800 and 6 ATI Readon
9250 (the memory is never overloaded)

Thanks for you hepl

ps : please forgive my bad english so far from my natural language

 

[pr]

I understand that it s a very rare question that I m ready to ask and
that only few people can actually test but a theoric explanation and
solution will be really grateful.

OK.

Here is my problem :

I have a computer with 3 video cards and 6 monitors plugged (for
business purpose). The resolution for each monitor is 1280x1024
On each of this monitor I have a IE browser (full screen) showing the
same html local page.

You're right. I can't duplicate that test environment.

This HTML is very simple containing a 2 <img> (640x480 for one and the
other is very small) , 15 div nested on 3 three levels only containing
static text.
The content of the first of these div (this one is not nested) is
refreshed every 1 s (by a settimeout) with innerHTML
document.getElementById("MyDiv").innerHTML= "Hello"

You need to be more specific.

window.setTimeout(functionObj, 1000)

is likely to perform better than

window.setTimeout("[long string of code]", 1000)

for example. Also if you literally want to write "Hello" (which contains
no markup) why use innerHTML? You will find it much quicker to change
the nodeValue property of a DOM text node instead. Please supply a short
but realistic example.

The HTML contains no javascript except the settimeout (started onload)
and is linked to a CSS. I don t know if it s important but the size of
every div is in % and the fontsize is em

That could be significant if the browser has to recalculate the position
of page elements every second. A fixed height and provision for overflow
in the 'updatable' div would seem to be a good idea.

The problem is that the process time used by each browser go to 11 %
and stay at this value, so a permanent total of 66 % for a simple
recurrent innerHTML (stay at this amount as long as the browsers are
running) on 6 browsers. I find this value very high and unlucky for me
non compatible for the use i want to made with my application.

I have tried other solutions like the well known "replaceHTML" or DOM
method without any success. The only way I have found to take down the
process time to 0 %!!! for every iexplorer.exe and keep my recurrent
innerHTML refresh, is to open 6 new browsers containing only one div
and this timer (and remove the div and the timer from the first 6
browser). I can conclude that the problem is caused by using a
recurrent innerHTML on my simple html page.

Do you think that's normal ? Have you any solutions or ideas about
that ?

Supply an example and somebody might have some.

[...]