Some months ago I purchased an Owon vds1022i usb oscilloscope.
It uses a Java application for control and display.

Testing it with an old Pi2 running Bookworm and the default GUI
showed very nice performance, with the screen updating the Owon
control application fast enough that no delay was noticeable.

Now, several months and Bookworm updates later, I tried it again
and the oscilloscope application updates the screen roughly once
every five seconds.

Top reports about 0.6% user with the Java application not running,
but close to 30% with the java application running but just sitting
in the foreground with no inputs.

Are there any known problems with Java on Bookworm? I've tried
disconecting other usb devices and swapping keyboards, that didn't
help.

Is it possible to "roll back" the system by date? The initial tests
that worked well were done in early August, 2024.

Thanks for reading,

bob prohaska

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Theo <theom+news@chiark.greenend.org.uk> wrote:

bp@www.zefox.net wrote:

Some months ago I purchased an Owon vds1022i usb oscilloscope.
It uses a Java application for control and display.

Java or Javascript? Javascript you'd run in a browser, Java you might download a .jar file and run it.

Not sure. The top window reports running java, but as I look at https://github.com/florentbr/OWON-VDS1022
there is a reference to python but no mention of java. Yet,
java is what uses most of the cpu when I start the vds1022i
control program. Is there some connection between the two?

Thanks for writing,

bob prohaska

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

(Apparently the question is about Java, not JavaScript.)

bp@www.zefox.net writes:

Some months ago I purchased an Owon vds1022i usb oscilloscope.
It uses a Java application for control and display.

Testing it with an old Pi2 running Bookworm and the default GUI
showed very nice performance, with the screen updating the Owon
control application fast enough that no delay was noticeable.

Now, several months and Bookworm updates later, I tried it again
and the oscilloscope application updates the screen roughly once
every five seconds.

Top reports about 0.6% user with the Java application not running,
but close to 30% with the java application running but just sitting
in the foreground with no inputs.

Are there any known problems with Java on Bookworm? I've tried
disconecting other usb devices and swapping keyboards, that didn't
help.

https://bugs.debian.org/cgi-bin/pkgreport.cgi?archive=0;dist=unstable;ordering=normal;repeatmerged=0;src=openjdk-17
would be the place to look for bug reports. I don’t see anything
reflecting performance issues.

At any rate, a single application performing poorly is just not enough
to conclude there’s a problem with the JRE. You’re more likely to be
seeing poor behavior by the one application.

Contacting the maintainer of the oscilloscope application would be a
reasonable approach (using whatever their preferred support or bug
reporting channel is).

Is it possible to "roll back" the system by date? The initial tests
that worked well were done in early August, 2024.

If you can find a copy of the JRE package from that time period then potentially yes. I’m not sure if the package binaries are kept anywhere
(in principle they can probably be rebuilt from source, if not, but this
is starting to turn into a lot of work for something speculative).

--
https://www.greenend.org.uk/rjk/

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 28/02/2025 15:22, bp@www.zefox.net wrote:

It appears the problem was Wayland. I was quite firmly convinced that
the initial experiment was done with Wayland active, but evidence
suggests otherwise.

Now that IS interesting. I would have expected that Wayland would be
faster...

X is such and enormous pile of bloated excrement that Wayland must be
very early stage design.

Switching from Wayland to X11 seems to have restored the former level
of performance. No other changes, I just set X11 active and rebooted.

I suppose it's possible there was a change to Wayland, quite a few
updates were applied between August and now. Alas, it's far more
likely I was simply mistaken about the initial test.

Thanks for everyone's attention, and apologies for the noise!

Nope. Its a really good nugget of info to tuck away.

--
Climate Change: Socialism wearing a lab coat.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Richard Kettlewell <invalid@invalid.invalid> wrote:

(Apparently the question is about Java, not JavaScript.)

bp@www.zefox.net writes:

Some months ago I purchased an Owon vds1022i usb oscilloscope.
It uses a Java application for control and display.

Testing it with an old Pi2 running Bookworm and the default GUI
showed very nice performance, with the screen updating the Owon
control application fast enough that no delay was noticeable.

Now, several months and Bookworm updates later, I tried it again
and the oscilloscope application updates the screen roughly once
every five seconds.

Top reports about 0.6% user with the Java application not running,
but close to 30% with the java application running but just sitting
in the foreground with no inputs.

Are there any known problems with Java on Bookworm? I've tried
disconecting other usb devices and swapping keyboards, that didn't
help.

https://bugs.debian.org/cgi-bin/pkgreport.cgi?archive=0;dist=unstable;ordering=normal;repeatmerged=0;src=openjdk-17
would be the place to look for bug reports. I don’t see anything
reflecting performance issues.

At any rate, a single application performing poorly is just not enough
to conclude there’s a problem with the JRE. You’re more likely to be seeing poor behavior by the one application.

Contacting the maintainer of the oscilloscope application would be a reasonable approach (using whatever their preferred support or bug
reporting channel is).

Is it possible to "roll back" the system by date? The initial tests
that worked well were done in early August, 2024.

If you can find a copy of the JRE package from that time period then potentially yes. I’m not sure if the package binaries are kept anywhere
(in principle they can probably be rebuilt from source, if not, but this
is starting to turn into a lot of work for something speculative).

It appears the problem was Wayland. I was quite firmly convinced that
the initial experiment was done with Wayland active, but evidence
suggests otherwise.

Switching from Wayland to X11 seems to have restored the former level
of performance. No other changes, I just set X11 active and rebooted.

I suppose it's possible there was a change to Wayland, quite a few
updates were applied between August and now. Alas, it's far more
likely I was simply mistaken about the initial test.

Thanks for everyone's attention, and apologies for the noise!

bob prohaska

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

bp@www.zefox.net wrote:

Some months ago I purchased an Owon vds1022i usb oscilloscope.
It uses a Java application for control and display.

Java or Javascript? Javascript you'd run in a browser, Java you might
download a .jar file and run it.

Theo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)