JDK 1.4 beta vs JDK 1.3.1 Performance?
Michael P. McCutcheon
I've heard mixed things about it on javalobby.org, some say it's faster and
uses less memory, some say it's slower.
Those new API's look really cool (logging, expressions, etc). Plus
assertions will be nice.
Cool!
Mike
Jon Skeet
> Anyone got any benchmarks?
>
> I've heard mixed things about it on javalobby.org, some say it's faster and
> uses less memory, some say it's slower.
The two small benchmarks I've done are almost identical. This is in no
way conclusive :)
> Those new API's look really cool (logging, expressions, etc). Plus
> assertions will be nice.
Indeed - today is "investigate new stuff" day :)
--
Jon Skeet - sk...@pobox.com
http://www.pobox.com/~skeet
If replying to the group, please don't mail me at the same time
Tim Tyler
: I've heard mixed things about it on javalobby.org, some say it's faster and
: uses less memory, some say it's slower.
My overall impression is that it's not bad.
However... if anyone wants some ammo to claim that Sun's current JVMs are
crap, I present my "HAL" applet.
You can see this running under your browser's native VM at:
The absolutely identical applet under Sun's plugin JVM can be seen at:
http://alife.co.uk/hal/plugin/
I estimate Sun's 1.4 JVM runs this program over 100 times slower than
Microsoft's JVM - and can fail miserably to redraw the screen :-|
--
__________
|im |yler t...@iname.com Home page: http://alife.co.uk/tim/
Tim Tyler
: However... if anyone wants some ammo to claim that Sun's current JVMs are
: crap, I present my "HAL" applet.
: You can see this running under your browser's native VM at:
: The absolutely identical applet under Sun's plugin JVM can be seen at:
: http://alife.co.uk/hal/plugin/
The problem is intermittent - I have now seen it running properly ;-)
It seems that in order to elicit the poor behaviour, it is necessary to
view some other applets first.
I seem to be able to reproduce the problem on request here.
When it fails it first prints a few error messages in the console.
It looks like it has problems shutting down a thread and then gets its
knickers permanently twisted - and won't run any more applets properly
until you shut it down.
FWIW, here's how things start:
java.lang.NullPointerException
at sun.plugin.ActivatorAppletContext.showStatus(Unknown Source)
at sun.applet.AppletPanel.showAppletStatus(Unknown Source)
at sun.applet.AppletPanel.run(Unknown Source)
at java.lang.Thread.run(Unknown Source)
java.lang.IllegalMonitorStateException: current thread not owner
at java.lang.Object.notifyAll(Native Method)
at halcell.stop(JAX)
at sun.applet.AppletPanel.run(Unknown Source)
at java.lang.Thread.run(Unknown Source)
netscape.javascript.JSException: Method invocation failed: 0x80004001
at sun.plugin.javascript.ocx.JSObject.nativeInvoke(Native Method)
at sun.plugin.javascript.ocx.JSObject.invoke(Unknown Source)
at sun.plugin.javascript.ocx.JSObject.setMember(Unknown Source)
at sun.plugin.ActivatorAppletContext.showStatus(Unknown Source)
at sun.plugin.AppletViewer.showAppletStatus(Unknown Source)
at sun.plugin.ocx.ActiveXAppletViewer.showAppletStatus(Unknown Source)
at sun.applet.AppletPanel.runLoader(Unknown Source)
at sun.applet.AppletPanel.run(Unknown Source)
at java.lang.Thread.run(Unknown Source)
Tim Tyler
: It looks like it has problems shutting down a thread and then gets its
: knickers permanently twisted - and won't run any more applets properly
: until you shut it down.
On further inspection that doesn't seem to be the (only) problem.
I've now seen this behaviour from a cold start, with the following error
coming from the applet on the first page looked at:
netscape.javascript.JSException: Method invocation failed: 0x80004001
at sun.plugin.javascript.ocx.JSObject.nativeInvoke(Native Method)
at sun.plugin.javascript.ocx.JSObject.invoke(Unknown Source)
at sun.plugin.javascript.ocx.JSObject.setMember(Unknown Source)
at sun.plugin.ActivatorAppletContext.showStatus(Unknown Source)
at sun.plugin.AppletViewer.println(Unknown Source)
at sun.plugin.usability.Trace.firePrintlnEvent(Unknown Source)
at sun.plugin.usability.Trace.netPrintln(Unknown Source) [...]
From the "usability.Trace" I would guess it seems likely that I am looking
at some debugging code here.
Anyway - to try to end on a positive note - /most/ of the time the new
beta works very well for me. My simple timings indicate that this is the
fastest Sun JVM on my code by quite a stretch - and that Java 2 has
finally started beating the old 1.1.x JVMs on intensive graphical tasks ;-)
--
__________
|im |yler There's no escaping the Rockz - http://rockz.co.uk/
Ronggolawe
Thanks
In article <9eht2o$5lr$1...@news.value.net>, "Michael P. McCutcheon"
Jon Skeet
> Where exactly I can download jdk1.4 beta ?