Use ZGC for garbage collection

[Sheikah45]

On my system noticed the OS reported memory usage was halved

[codecov]

Codecov Report

Merging #2951 (650f03c) into develop (b539935) will decrease coverage by 0.01%.
The diff coverage is n/a.

❗ Current head 650f03c differs from pull request most recent head cbcf770. Consider uploading reports for the commit cbcf770 to get more accurate results

Additional details and impacted files

@@              Coverage Diff              @@
##             develop    #2951      +/-   ##
=============================================
- Coverage      61.34%   61.33%   -0.01%     
  Complexity      4595     4595              
=============================================
  Files            557      557              
  Lines          20186    20186              
  Branches        1048     1048              
=============================================
- Hits           12383    12381       -2     
- Misses          7217     7219       +2     
  Partials         586      586              

see 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b539935...cbcf770. Read the comment docs.

[BlackYps]

Halved? That's huge!

[magge-faf]

Yeah, that is amazing - have found no bugs so far, still testing it.

[Sheikah45]

Yeah I ran through some tests with it. Still not entirely sure why. A lot of the memory used by the client is actually off heap, so it is rather hard to track down. So I was a little surprised when using the ZGC collector reported so much less memory usage. I am still investigating some as to why hence why this is not merged but figured I would send it out to be tested so others could confirm if it happened for them as well.

[Sheikah45]

And previously this "unknown" memory segment was partly due to the Webviews. But using the ZGC the os reported memory actually correlates with the heap+nonheap memory. Where as before ~500MB were due to either the GC or some other native memory purpose. But I do not know why anything other than the memory of the GC would change just from the GC change.

[micheljung]

Why?

[Sheikah45]

With how ZGC operates it needs a little bit more heap overhead in order to operate properly without constantly performing concurrent GCs and kicking up CPU usage

[micheljung]

Comment would also need to be adjusted (or just removed).