dev@grizzly.java.net

Re: Tutorial Contribution

From: Ken Cavanaugh <Ken.Cavanaugh_at_Sun.COM>
Date: Mon, 23 Jul 2007 15:19:21 -0700

Sven Reimers wrote:
> Finally - enjoy the bugs ;-)
I will :-).

Is it possible to filter the output to ignore certain packages?
For example, there is some third-party code in the ORB, as well as
some "known bad code" packages that I'd like to selectively ignore at times.

Thanks,

Ken.

>
> -Sven
>
> On 7/23/07, charlie hunt <charlie.hunt_at_sun.com> wrote:
>> Ken Cavanaugh wrote:
>> > charlie hunt wrote:
>> >
>> >> This is indeed very puzzling >:o
>> >>
>> >> I just grabbed the update from the latest (internal) CORBA workspace.
>> >>
>> >> I modified my netbeans.conf by quickly just changing previous set
>> >> value of -J-Xmx384m to -J-Xmx1g.
>> >>
>> >> I did a Clean and Build of the project since I had new updates
>> >> brought into my workspace.
>> >>
>> >> Then, I ran FindBugs. It completed in about 2 minutes. My Java heap
>> >> is currently at 660m. You can display a GC monitor icon in NetBeans
>> >> by doing View > Toolbar and checking the "Memory" box.
>> > My NB instance won't let me do that at present.
>> > I'll restart and try again (that worked).
>> >>
>> >> Btw, you can also add a -J-verbose:gc to your netbeans.conf JVM
>> >> command line and then launch netbeans from the command line. The
>> >> output from the -verbose:gc will go to stdout / stderr.
>> >>
>> > I was using a 64 bit JDK 5 Linux VM. I've changed to 32 bit JDK 6
>> > Linux VM. My peak heap
>> > usage is coming close to 800 MB, but a full GC dropped that
>> > substantially.
>> >
>> > It worked this time! The problem may have been caused by using a 64
>> > bit JVM on Linux.
>> > That's the main thing I changed.
>> >
>> > Thanks,
>> >
>> > Ken.
>>
>> Yes, a 64-bit JVM will require a larger Java heap as a result of the
>> larger pointer / object references.
>>
>> It's likely that 768m value was right at the point of no return with a
>> 64-bit JVM.
>>
>> Glad to hear it is working!
>>
>> charlie ...
>>
>> --
>>
>> Charlie Hunt
>> Java Performance Engineer
>> 630.285.7708 x47708 (Internal)
>>
>> <http://java.sun.com/docs/performance/>
>>
>>
>>