If your program runs on a JVM, you can test in pre...
# nucleus
u
If your program runs on a JVM, you can test in preview in the last
2.6.0-dev-202609241732
Copy code
nucleus.application {
    nucleusOptimization = true
}
it will automatically apply several optimizations so that your program starts much faster, especially on Windows, and consumes much less memory, well obviously, it is strongly advised in addition to enabling aot cache as well as proguard
The text on the image is incorrect, JDK 27 is stable today, and it's not just about having the latest JDK, the gains are much more significant.
p
I'm using Nucleus as dependency (not applying plugin). On that build it consumes after some time app run about 1.1G of memory. Most of it is seems like native memory. Is there any configuration I can tune to decrease native memory usage?
u
This did not happen before on Nucleus 2.5 ?
p
I was using AWT backend before. One moment I'll check
With previous version (2.5.18) with AWT backend it is the same 1.1G. Java is using Xmx256m with 90m metaspace Other seems to be native memory
u
Can you profile the application that we see, where does it really come from? Are there many animations?
p
Noticed that when making lot of animations it is increasing rapidly till about 1G, then slowly till 1.1G
u
Maybe relative to github.com/NucleusFramework/Nucleus/issues/541 it would be interesting to try applying the nucleus plugin with the optimizations despite everything
p
Now I try to relaunch several times just with minimized window. It is initially around 800mb. I'll try later to find out where it happens (I'm not expert in native code profiling, it will take time)
u
if you have intellig ultimate, you could use the profiler, we would see very quickly where it comes from
p
I'm on CE. Will extract repro code and if it will be consuming same amount of memory then share with you.
👍 1
I've checked on another PC with Windows and screen resolution 1920x1080 - it is around 400mb total app memory usage
On Linux PC with resolution 3200x2000 it is till 1.1G. Just thinking may be resolution is the cause of higher memory consumtion
I'll check tomorrow that assumption