Skip to main content

6u14-b01 is available

16 replies [Last post]
rogyeu
Offline
Joined: 2006-07-30
Points: 0

Hi,

The early access build for Java SE 6 Update 14 build 01 is now available. Grab it if you haven't already. Major features in this release includes:
- Service Tag creation for Windows JRE
- Java HotSpot 14

For more info, please see https://jdk6.dev.java.net/6uNea.html

Thanks,
-- RY

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
rogyeu
Offline
Joined: 2006-07-30
Points: 0

We putback 6784816 late in the previous release and have missed the window for putting it back in 6u14-b01. We plan to include the fix in 6u14-b02.

jacobdk
Offline
Joined: 2008-04-21
Points: 0

Thanks Roger - that's good to hear. Just please make sure it's included it in 6u13 too - but I doubt that should be a problem. :-)

Keep up the good work. Looking forward to what's coming.

Jacob

rogyeu
Offline
Joined: 2006-07-30
Points: 0

Yes. They'll be in 6u13.

p.s. i need to get back to you about jdk7 (if i have more public info).

rogyeu
Offline
Joined: 2006-07-30
Points: 0

G1 is in Beta right now. It is not enabled by default. The parallel collector is still the default GC and is the most efficient GC for common household usage. G1 is meant to be an alternative for the concurrent collector. G1 is suppose to be more predictable and enable fast allocation with memory regions design.

To use it: -XX:+UnlockExperimentalVMOptions -XX:+UseG1GC

For more information about GC & G1, please see:
G1 Technical Session in 2008 JavaOne: http://developers.sun.com/learning/javaoneonline/j1sessn.jsp?sessn=TS-54...
Java SE 6 GC tunning: http://java.sun.com/javase/technologies/hotspot/gc/gc_tuning_6.html

If you want more than above, here's a research paper with the analysis:
http://research.sun.com/jtech/pubs/04-g1-paper-ismm.pdf

Please keep in mind some of the configurations are for big applications which requires lots of threads and memory management. Performance tuning usually focuses on a specific case and environment; therefore, one solution might not fit other cases/apps.

6u14 includes performance enhancements in core functions and HotSpot. I am glad if you can recognize the differences. Thank you for trying the 6u14 ea builds.

-- RY

demonduck
Offline
Joined: 2008-03-14
Points: 0

Thanks for more complete information on G1. I don't expect to see any difference in performance in my applet but I'm going to take a look. And because I use an old, slow machine running Windows 2000, performance times are magnified. It's like watching in slow motion.

And it will be reassuring just to know it works with my applet on my machine. Who knows, I might see something unusual....

After tests:

I saw something unusual. Nothing to do with G1.

It only happens in Plugin2. I have FF2 and IE6. FF2 does not use the new plugin so I don't see this in FF2. IE6 does use the new plugin and I see the problem in IE6 so this is clearly a Plugin2 problem.

Try this applet:
http://pancyl.com/VSync.htm

Notice the frame rate. Now click inside the applet to give it focus. Notice that the frame rate slows. Now click outside the applet to blur and notice that the applet speeds up again.

Do you see it? Or is this just something that is happening on my machine.

The good news is that 6u14b01 is more than twice as fast as 6u10 on the same applet. That's really good news.

But this focus thing is weird....

suryad
Offline
Joined: 2005-02-23
Points: 0

TBH I tried it and I did not notice any difference. I noticed the fps was over 2000 almost 3000 on my machine regardless of whether I click outside or inside.

Hope that helps! :)

demonduck
Offline
Joined: 2008-03-14
Points: 0

> TBH I tried it and I did not notice any difference. I
> noticed the fps was over 2000 almost 3000 on my
> machine regardless of whether I click outside or
> inside.
>
> Hope that helps! :)

Must be just a quirk of my old machine because it's so old. Current generation of desktops are so fast now that what ever I'm seeing on my machine simply doesn't show. For example, I get about 30 fps. Your machine is a hundred times faster than my machine.

Thanks for trying it...

rgibson_2
Offline
Joined: 2006-02-28
Points: 0

Hi Roger,
I can't find any documentation about HotSpot 14 or the work done on TreeMap/HashMap, can you point me to some? For example, are compressed OOPs enabled by default? http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6732819 suggests not, but ...

Also, what is the relationship between Java 6 Performance Release 6 and this update? Is it the "upcoming standard release" mentioned at http://java.sun.com/javase/technologies/performance.jsp ?

Thanks for any pointers you can give,
Robert

rogyeu
Offline
Joined: 2006-07-30
Points: 0

Hi Robert,

There aren't any API changes in 6u14; therefore, there's no documentation. However, when 6u14 is released, info (ie. compressed OOP) should be included in release notes. Do you have an specific questions (please start a new thread)?

6 Update 6 Performance Release uses HotSpot 13 plus 6 Update 6 core lib. 6u14 has the latest core lib + HS14.

HTH.
RY

jacobdk
Offline
Joined: 2008-04-21
Points: 0

Just a few questions for clarification.

I'm just wondering, why your next beta is 6u14 and not 6u13? Will it be like this in the future, that you will release each second update as a public beta, like you did with 6u10 and 6u12? The releases inbetween will only be security and other minor fixes - as I remember, 6u11 was only minor security updates, am I right?

Is this a practise, you plan to continue with, when you release Java 7 sometime? Ie. release each second update as a public beta?

Edit: as I posted in another thread too, bug 6784816 has returned in this build. Please fix it asap....

Thanks in advance,
Jacob

tackline
Offline
Joined: 2003-06-19
Points: 0

There's an answer here: http://weblogs.java.net/blog/opinali/archive/2009/02/here_comes_jdk.html

In fact, update numbers are subject to change (which can cause confusion when something that should be appearing in 6u(n) appears in 6u(n+1) because of an inserted update).

ivan_g_s
Offline
Joined: 2008-08-23
Points: 0

Wow! The new GC is really much better.
Doing a certain benchmark to compare the old and new GCs,
I really experienced that G1 doesn't "stop the world".

I was testing an intepreter. It allocates a lot of short-lived objects,
when testing recursions.
I tested it several times as I couldn't believe my eyes.
The benchmark results using G1 are so homogeneous (and
also a lot faster).

Look at the difference:

"usual GC":

recursion2 #1 3.51s
recursion2 #2 1.21s
recursion2 #3 1.61s
recursion2 #4 1.59s
recursion2 #5 1.39s
recursion2 #6 1.96s
recursion2 #7 1.39s
recursion2 #8 1.82s
recursion2 #9 1.67s
recursion2 #10 1.11s

time: 1.11s < 1.72s < 3.51s (min < average < max)
recursions: 7500

"G1 GC":

recursion2 #1 1.3s
recursion2 #2 1.19s
recursion2 #3 1.03s
recursion2 #4 1.06s
recursion2 #5 1.05s
recursion2 #6 1.04s
recursion2 #7 1.0s
recursion2 #8 1.11s
recursion2 #9 1.09s
recursion2 #10 1.05s

time: 1.0s < 1.09s < 1.3s (min < average < max)
recursions: 7500

cowwoc
Offline
Joined: 2003-08-24
Points: 0

What is the status of the G1 collector? Is it enabled by default? If not, how does one enable it?

Thank you,
Gili

demonduck
Offline
Joined: 2008-03-14
Points: 0

We know that G1 exists in 6u12b01:

6622418 3-Medium Defect G1: assert(false,"Non-balanced monitor enter/exit!") fails
6652160 3-Medium Defect G1: assert(cur_used_bytes == _g1-\>recalculate_used(),"It should!") at g1CollectorPolicy.cpp:1425
6660573 3-Medium Defect G1 : BigApps Failure : guarantee(satb_mq_set.completed_buffers_num() == 0,"invariant")

---------

We know how to enable the use of experimental XX JVM args:

6618726 3-Medium RFE Introduce -XX:+UnlockExperimentalVMOptions flag

--------

There is a paper that describes and alludes to G1:

http://blogs.sun.com/jonthecollector/entry/our_collectors

--------

So why the guessing games? Just tell us...

mbien
Offline
Joined: 2007-04-29
Points: 0

good work. Everything seems to run smoother, i noticed no regression so far. This are lot of bugfixes and improvements for a single build ;)

demonduck
Offline
Joined: 2008-03-14
Points: 0

Seems to run my applets a little faster. I didn't do a timing test but it seems crisper. Maybe the new HotSpot???

But this still occurs:

http://forums.java.net/jive/thread.jspa?threadID=55700&tstart=15