Java Arg help
crashing with
-XX:+UnlockExperimentalVMOptions -XX:+UseZGC -XX:+ZGenerational java arguments, heres the not really a log log: https://mclo.gs/0300X9m57 Replies
im using prism and putting them here if that matters
and it runs with the same arguments removed right?
Unrecognized VM option 'ZGenerational'
Process exited with code 1.
i'm not seeing anything wrong but i could be missing something obvious
it runs fine without those arguments yea
ah zgc generational on java 17 probably isn't supported
can you try installing adoptium 21?
prism lets you disable the compatibility check
yeah, i can then just force disable compat check
yea alr
https://adoptium.net/ the download on the main page here right
Adoptium
Adoptium
Eclipse Adoptium provides prebuilt OpenJDK binaries from a fully open source set of build scripts and infrastructure.
it is, but prism has this built in, unless you specifically want to install it systemwide
oh it does
for some reason the "download java" button isn't present in global settings
edit your instance
ohh i can download it thru the management tab
still gives me that pop up error https://mclo.gs/8ICqsx1
🤔
oh
i had to select it per instance and not globally
oh you said thatðŸ˜
im still getting these tps spikes tho even with those args
can you send your new latest.log?
interestingly it removed the fps spikes from earlier but the internal server is still dying. as a sanity check what if you set dh's cpu load preset to minimum impact?
was there fps spikes?
oh yea a tiny bit ig
ill try minimum impact
very subtle ones but they were there
yea
and they were at regular intervals which is typical of gc stutters so it makes sense for those to be gone. though the same could be said for the internal server spikes
those staying is strange
this is minimal impact
i was on low impact before
or whatever the one right after minimal is
that's low yeah
log is really plain too, nothing sticks out
im trying 1.21 neoforge rn just to see
how do you enable pie charts on 1.21 lol
or the tps/fps graphs i mean
f3 and 1 2 or 3
my 60% board was not built for this
yeesh yea well its not bc im on 1.20.1 ig
seems much worse on 1.21 for some reason
uhh ima try rc1
it's going off the screen. absolutely wonderful
is this not rc1? https://discord.com/channels/881614130614767666/882038774861819954/1014726512420868096
its showing as 2.3.3-b-dev
those are nightly builds, but rc1 released recently so there's only tiny differences anyway
also what's your total system memory usage look like? considering you went down in allocation by 24gb i doubt it's hitting swap, but it's best to have it ruled out
ill check once i can load in, but did internal server get renamed to "Full - Save Chunks" ?
yup
looks like my system memory sits around 26gb when im in the world, and 2.3.3-b-dev didn't seem to solve it either, this is with internal server, the java args you sent and lz4 compression
what if you add c2me to at least multithread the worldgen (assuming that's what's causing spikes, it typically wouldn't though)
i can try for this test but i can't use c2me and distant horizons with my normal pack because i use terrafirmacraft: https://discord.com/channels/881614130614767666/1388627785492332675
but ill try rn
that does seem to fix it though
now i just need to figure out how to use DH + C2ME alongside terrafirmacraft
which is what i would prefer to do anyway
bc that gen speed is crazy
best course of action would probably be to leave it on overnight at the highest cpu load preset
when you're back next morning disable distant generation entirely
can you see anything wrong with tfc + dh + c2me though? id rather just be able to use all 3
if you find a way to have them work together it should be fine
https://mclo.gs/o3684lo i get this crazy log spam with all 3
results in some holes in the lod and being unable to save the world, and eventually crashes after a bit
everything runs fine with c2me alone: https://mclo.gs/LYIpQ7m
and with dh alone, but as soon as i add all 3 it explodes
try removing dh, using chunky to generate it all with c2me as well, removing chunky and adding dh back
i mean im sure that would work but is there no solution that involves using all 3
not at once out of the box, this'll take a bit longer but you'd get the result you're looking for
its just with tfc you kinda gotta explore long distances some times so its a bit unreasonable to use chunky in that way
ig im just wondering if this is something dh needs to fix, c2me needs to fix or tfc needs to fix
from a quick glance at the log i can't really tell
lot of NullPointerException
ill try asking c2me peeps