144531 - java applets are EXTREMELY slow.

Status: RESOLVED WORKSFORME

(Core Graveyard :: Java: OJI, defect)

Description

[Zurk]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.0rc2) Gecko/20020510
BuildID:    2002051006

There is a noticeable performance lag compared to
netscape 4.78 and the 0.9.7 mozilla builds.
This is with jdk 1.4 and 1.3.1..applets run at
least 10x slower than with either netscape 4.78 
or mozilla 0.9.7. This is especially noticeable
with multithreaded applets. The same JVMs & plugins
on the other browsers run much faster.

Reproducible: Always
Steps to Reproduce:
Run a compute intensive applet and watch how
sluggish it is. note that im using SCRIPTABLE
and MAYSCRIPT enabled  in the applet tag which 
may also have an effect.

Actual Results:  very slow performance. between 10x and 20x slower.
especially with threaded applets. some of my applets
take nearly 15 MINUTES to run on a P133 vs 30 seconds
for the same applet on netscape 4.78 on the same 
machine with identical JVMs. JDK 1.4 helps slightly as
does disabling talkback on mozilla but not by much.

Expected Results:  faster performance please!

Comment 1

[Olivier Cahagne]

Sev -> Normal.

Comment 2

[Stuart Lamble]

I'm seeing something similar. I have Mozilla 1.1, running on Mac OS X on a white
iBook (G3 500, 256 MB RAM). Where I work is using a Bridges Java client to run a
call-tracking package. If I open the client in IE, it runs at a reasonable pace
(although there are other problems with the screen drawing). If, however, I open
it with Mozilla, it is like molasses.

The really interesting thing is that I see the same performance in IE if Mozilla
is up and running. If I shut down Mozilla, performance jumps up drastically.

I'd post a relevant URL here, but I'm not sure if it can be accessed from
outside Monash, and even if it can, I'm not certain that it would be useful, as
the obvious performance problems occur once I'm logged into the application (not
at the login screen, although there _may_ be some manifestations...)

Comment 3

[Dennis Daniels]

need a URL so that we can all test this and get a benchmark...
http://members.aol.com/shadows125/connect4.htm
we'll try that one.
It wfm 
Mozilla 1.2b
Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.2b) Gecko/20021016
Java(TM) Plug-in: Version 1.4.1
Using JRE version 1.4.1 Java HotSpot(TM) Client VM
User home directory = C:\WINDOWS

Proxy Configuration: Browser Proxy Configuration

Comment 4

[Stefan A. Möller]

I see this problem, too. Here's a nice testcase:
http://www.heise.de/ct/browsercheck/java.shtml
You can play with that Rubik's Cube using CTRL + left mouse button.
Whenever you make a move there's a 2 seconds delay until the java applet reacts.
Other browsers that are using the JRE directly (without plugin), like Opera or
Konqueror, are very much faster here. Maybe we have the problem that only the
plugin is so slow.

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021212
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.1_01-b01)
Java HotSpot(TM) Client VM (build 1.4.1_01-b01, mixed mode)

Comment 5

[tmh]

This is a really bad bug - Mozilla 1.2 was fine but Mozilla 1.3 is basically
unusable on any page with java in it.  For an example see http://news.bbc.co.uk
- the applet on this page causes Mozilla to use 100% CPU and stop responding. 
In IE with exactly the same java plugin (1.3.1) (and mozilla 1.2 with the same
plugin) this page is fine.

Comment 6

[Jan Van Looy]

this is simply awful. on IE with the very same VM it works fine. so much for ms
discriminating java :)

http://agents.www.media.mit.edu/groups/el/projects/spacewar/

Comment 7

[Michael Gabriel]

jre 1.4.2, moz 1.5b,on win2k all working fine
http://www.heise.de/security/dienste/browsercheck/tests/java.shtml
and
http://agents.www.media.mit.edu/groups/el/projects/spacewar/
and
http://members.aol.com/shadows125/connect4.htm
all np

Comment 8

[Jan Van Looy]

moz 1.5b (&firebird) appears ok. status fixed?

Comment 9

[Stefan A. Möller]

I just discovered that in my case the problem originates in the artsdsp sound
redirection. Starting Mozilla with "mozilla" all Java applets run fast, when
started with "artsdsp mozilla" some Java applets (e. g.
http://heise.de/security/dienste/browsercheck/tests/java.shtml) are painfully
slow (and without sound anyhow).

Comment 10

[Andrew Hagen]

Please try Mozilla 1.7 RC1 or later. Does the problem still occur?

Comment 11

[Kenan]

http://heise.de/security/dienste/browsercheck/tests/java.shtml

WFM with Build 2004061308 and WinXP & JRE 1.4

Comment 12

[Jan Van Looy]

This bug has been fixed in more recent builds. Could anyone change the status to
resolved?