[prev in list] [next in list] [prev in thread] [next in thread]
List: kfm-devel
Subject: updated rendering speed tests [testcases attached]
From: Vadim Plessky <lucy-ples () mtu-net ! ru>
Date: 2001-11-26 11:45:34
[Download RAW message or body]
On Sunday 25 November 2001 00:33, Thomas Schilling wrote:
| > (smaller number means better result)
| > Konqueror Mozilla
|
| Opera5/Lin
|
| > LongHTML-P 848-934 1045-1063 4319-11804
| > LongHTML-DIV 641-850 1022-1027 11432-11706
| > LongHTML-DIV-CSS 1334-1346 1063-1097 4532-11835
|
| Something to despair ...
|
| IE (PIII 667Mhz 256MB Win98)
| LongHTML-P 50-110 (mostly 60)
| LongHTML-DIV 50-110 (mostly 60)
| LongHTML-DIV-CSS 50-110 (mostly 60)
|
| Hopefully it's a measuring error ... ;)
Hi Thomas,
Unfortunately, this is *not* measurement error.
MS IE6 (and MS IE5, IIRC) is much faster on this code - as you can see, appx.
8-10 times faster than Konq or Mozilla.
hat is really funny, IMO, is that Opera "the fastest browser on the earth!"
is VERY SLOW on this code - in fact, slower than Netscape Navigator 4.7!
(you will wait 16sec. until Opera finished...)
You convinienced me to make testing under Windows [98] as well, so here is
update/add-on to previous results.
Windows 98, P III/600, 128MB
IE6/Win NS6/Win NN 4.7 Opera5/W
P 60-110 * 270-330 6590-6750 16540
DIV 110 270-280 6530-7080 16470-16540
DIV-CSS 110 280-330 6760-7420 16480-16540
TABLE-1cell 110 280-330 6430-7470 16530-16550
TABLE-CSS 110-160 270-280 6260-7080 16530-16970
TABLE-CSS-2cell 160-170 330 6700-7690 16530-16590
* time measured in ms.
Updated results for Linux
LM 8.0, KDE 2.2.1, P III/600, 128MB
Konqueror Mozilla Opera5/Lin
LongHTML-P 848-934* 1045-1063 4319-11804
LongHTML-DIV 641-850 1022-1027 11432-11706
LongHTML-DIV-CSS 1334-1346 1063-1097 4532-11835
TABLE-1cell 1769 1622 4890-5253
TABLE-CSS 2179-2242 1611 4972-5241
TABLE-CSS-2cell 3732 1878 4984-5270
* time measured in ms.
If you compare results for NS6/Win (which IIRC was using Mozilla M18
milestone) and Mozilla 0.9.5/Linux - you will see that NS on Windows is
faster than rescent Mozilla on Linux (and faster than Konq). I have no
explanation to it excpet there was regression in recent Mozilla builds (or
Mozilla is slower on Linux comparing to Windows)
On the other hand, Opera/Linux is almost 3 times faster than Opera/Windows
(which really wonders me, as Opera was claming common database for all their
platforms).
Summary:
~ ~ ~ ~ ~ ~
I don't know wether 60-110ms MS IE has is an achievable target. But I guess
Konq should be in NS6/Win range (270-330ms).
Anyway, rendering speed of KHTML is not slow exactly on Tables (if you can
call it "slow" at all)- either P, DIV or TABLE/TD have similar rendering
speed results
I attach to this message TABLE testcases, which complement previous P, DIV
and DIV-CSS.
|
| Thomas
--
Vadim Plessky
http://kde2.newmail.ru (English)
33 Window Decorations and 6 Widget Styles for KDE
http://kde2.newmail.ru/kde_themes.html
KDE mini-Themes
http://kde2.newmail.ru/themes/
["LongHTML_TABLE-1cell.html" (text/html)]
<HTML>
<HEAD><TITLE>Test for Long HTML files</TITLE>
<style type="text/css">
DIV { width: 150px; height: 60px;
border: 1px solid lime;
margin: 10px;
}
</style>
</HEAD>
<BODY>
<H1>Long HTML: 600 <TR><TD> ... </TD></TR>tags with text inside</H1>
<SCRIPT language=JavaScript>
<!--
timestamp = new Date ();
var sec1, sec2, ms1, ms2, ms, tim1, tim2, timdif;
tim1 = (timestamp.getTime());
sec1 = timestamp.getSeconds();
ms1 = timestamp.getMilliseconds();
document.writeln("<TABLE>");
for (i=0; i < 600 ; i++)
{
document.writeln("<TR><TD>Text: line (cell) no " + i + "</TD></TR>");
}
document.writeln("</TABLE>");
timestamp = new Date ();
tim2 = (timestamp.getTime());
timdif = tim2-tim1;
sec2 = timestamp.getSeconds();
ms2 = timestamp.getMilliseconds();
if (sec1=sec2)
{ ms = ms2-ms1; }
else
{ ms = (sec2-sec1)*1000 + ms2 -ms1; };
// alert("sec1 ="+sec1+" sec2= "+sec2);
alert("time difference: "+timdif+"; - ms: "+ms);
//-->
</SCRIPT>
</BODY>
</HTML>
["LongHTML_TABLE-CSS.html" (text/html)]
<HTML>
<HEAD><TITLE>Test for Long HTML files</TITLE>
<style type="text/css">
DIV { width: 150px; height: 60px;
border: 1px solid lime;
margin: 10px;
}
table { width: 100% }
td { color: blue;
border: 1px solid lime; }
</style>
</HEAD>
<BODY>
<H1>Long HTML: 600 <TR><TD> ... </TD></TR>tags with text inside</H1>
<SCRIPT language=JavaScript>
<!--
timestamp = new Date ();
var sec1, sec2, ms1, ms2, ms, tim1, tim2, timdif;
tim1 = (timestamp.getTime());
sec1 = timestamp.getSeconds();
ms1 = timestamp.getMilliseconds();
document.writeln("<TABLE>");
for (i=0; i < 600 ; i++)
{
document.writeln("<TR><TD>Text: line (cell) no " + i + "</TD></TR>");
}
document.writeln("</TABLE>");
timestamp = new Date ();
tim2 = (timestamp.getTime());
timdif = tim2-tim1;
sec2 = timestamp.getSeconds();
ms2 = timestamp.getMilliseconds();
if (sec1=sec2)
{ ms = ms2-ms1; }
else
{ ms = (sec2-sec1)*1000 + ms2 -ms1; };
// alert("sec1 ="+sec1+" sec2= "+sec2);
alert("time difference: "+timdif+"; - ms: "+ms);
//-->
</SCRIPT>
</BODY>
</HTML>
["LongHTML_TABLE-CSS-2cell.html" (text/html)]
<HTML>
<HEAD><TITLE>Test for Long HTML files</TITLE>
<style type="text/css">
DIV { width: 150px; height: 60px;
border: 1px solid lime;
margin: 10px;
}
table { width: 100% }
td { color: blue;
border: 1px solid lime; }
</style>
</HEAD>
<BODY>
<H1>Long HTML: 600 <TR><TD> ... </TD></TR>tags with text inside</H1>
<SCRIPT language=JavaScript>
<!--
timestamp = new Date ();
var sec1, sec2, ms1, ms2, ms, tim1, tim2, timdif;
tim1 = (timestamp.getTime());
sec1 = timestamp.getSeconds();
ms1 = timestamp.getMilliseconds();
document.writeln("<TABLE>");
for (i=0; i < 600 ; i++)
{
document.writeln("<TR><TD>Text: line (cell) no " + i + "</TD><TD>2nd cell at right side</TD></TR>");
}
document.writeln("</TABLE>");
timestamp = new Date ();
tim2 = (timestamp.getTime());
timdif = tim2-tim1;
sec2 = timestamp.getSeconds();
ms2 = timestamp.getMilliseconds();
if (sec1=sec2)
{ ms = ms2-ms1; }
else
{ ms = (sec2-sec1)*1000 + ms2 -ms1; };
// alert("sec1 ="+sec1+" sec2= "+sec2);
alert("time difference: "+timdif+"; - ms: "+ms);
//-->
</SCRIPT>
</BODY>
</HTML>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic