Is this intended behavior? DH pregen process crawls to a half within minutes of genning.
I understand about "slowgen" and that it takes a while, but this is like decimal percentage of a single percentage. I attempted this and other testing compared to say chunky to DH's gen and there were no issues there. So it's only DH that's bricking like this?
62 Replies
do you have any idea how much 1500 chunks is
yea, thatll take a while
but the real problem is probably that you tried chunky with a radius of 1500, and that went a lot faster, right?
the reason chunky was a lot faster is because chunky uses blocks, so when you generated a radius of 1500 with chunky, it was a radius of 1500 blocks, which is 94 chunks
a square with a radius of 1500 is many many times larger than one with 94 chunks
a DH 1500 radius is 24000 blocks
Idk how much it changed but when genning with chunky previously I didn't do 1500 blocks, I did it with -c added at the end which changed the radius from 1500 to 24000. Which progressed through the night at a constant rate of about 33 chunks per second
What cpu preset do you have selected for DH?
I paid for the whole CPU I believe, I can double check
CPU sits at 16% usage and ram usage 11GB/16.
so with the i paid for the whole cpu preset, its using only 16% of the cpu?
that doesnt sound right
Went and grabbed food and came back after resetting back to DH pregen
Im confused as to how its full data when I have a 100+ SSD for the stroage usage as well
Are you using the DH features mode or internal?
Thats possible whats doing it because I believe it was set to internal which I know is much more intenseive but its the only way to accurately have everything placed right?
Yes
Internal is limited by the vanilla chunk generation
So you need to install C2ME and configure it to your thread count for the generation to speed up
Modrinth
Concurrent Chunk Management Engine (Fabric) - Minecraft Mod
A Fabric mod designed to improve the chunk performance of Minecraft.
Okay I do have c2me Installed but I dont know how to configure the thread count for c2me, I assume they may have explanations in their page so I shall check there and then check back?
Set the
globalExecutorParallelism to the thread count in the c2me config file in your config folder
The default is 8, which would equal to 4 coresI believe I found it
so default to 4?
globalExecutorParallelism
Default is 8 threads, which is 4 cores
You need to set it to whatever thread count you want to use
One core has 2 threads
The processor for my current plan is a 16 vCore CPU I'd prefer to dedicate like a quarter/to probably half to DH generation/pregen in hopes that so much will be genned the server won't have to compensate when it's already generated
Set it to 8 or 16
Okay 16 has been set and I'll attempt to run pregen again
Don't forget to restart the server
I did, just a simple on/wait till close/back on again I assume would have sufficed
Config has been saved and 16 is definietely set however console reports indicates the same issue with CPU usage
Gen mode IS set to Internal which I understand would take quite a while, so if thats it I understand
C2ME should increase the world gen speed quite a lot
It's seems to be sturggling hard, only increasing in increments of about .5 or so?
Weird
As for big world generation esque mods, I have BOP and BWG, Terrablender, and Continents.
I do have maybe 3-4 dungeon generation mods as well like When Dungeons Arise/Seven Seas as well Dungeons and Taverns, Yungs better dungeons, adventure dungeons, and repurposed structures, stuctory. Yet I decided to install sparse structures in hopes it would spawn everything less frequently
but even after doing that I ran into this same issue, but whenever I switch to try and pregen to chunky it works fine, but it also stalls out with DH when trying to load those chunky chunks into LODS. Which is whats happening now.
Since writing this the generation has ticked to 269.6/1500 so only 4 chunks?
if i'm understanding the ratio correct I believe
I constantly see it saying that Fulldata v2 issue with phantom recovery after about every generation.
I think it is EVERY small bit of progress actually
I believe if I leave it running pregenning like this, it does eventually run out of memory and both ram usage/CPU increase exponentially almost like a memory leak? I'll leave it running for now to see if it happens to get a crash report.
It's doing about 1 chunk per minute I think
Since its been 10 minutes since the last screenshots and ive only increased by 10
no its doing way more than 1 chunk a minute, because 1500 is the radius, not the total amount of chunks
So then this is an astrometrically huge number of chunks then, and thats why it takes so long?
It's still only using about 16% of cpu right now for being maxed on it's setting too
Is there anyway for me to see how many chunks per second or minute it is generating?
Yes
I just realized thats the progress of the radius i think
Yes that's a problem, not sure why it's doing that
Not really apart from calculating it yourself
Would I also be correct in assuming that since this a radius, that the way it generates isn't symmetrical but in like a circular rotation of generation? so unless I let it go all the way through, then I'll just have a very long distance in 1 direction but not the rest?
This is about 2 hours later after a nap
still nothing on the CPU front which is very odd
If theres anything else I can send just lmk
Are you sure you have C2ME?
Double check the logs if C2ME gets loaded
It should be okay to stop the server while its pregenning right?
It def says its in there but lemme check remote connection as well
Yep I do have it in there
I'll restart to make sure its loading in console
Yep detected in on startup
Yes, it's intended
It doesn't slow down but displays the radius you can already use.
LODs are generated in rings around the set origin, and each new ring takes more sections to generate fully
Are there any known issues with stopping a pregen and restarting it later at different locations/and or not the same spots?
So I'm testing to see if I can find any difference in CPU usage (right now sitting at 17%) but I'm worried if I stop/restarting the server etc to test things or change things will I potentially run into issues restarting the pregens?
Should work fine if you stop and resume later
Idk if it make sense or not but I decided to do 500 for the nether instead of attempting 1500 (Which I completed about 700-800 I believe.) Anyhow, again for the genning it just feels like its slowing down to a crawl.
2 days ago when genning a radius of 1500 in the OW, it genned about 300 in 3 hours. Rn it's struggling a smaller of fraction of the for the Nether with only Better Nether installed rn.
Its fine if this is again normal behavior, but it just feeeeels like somethings slightly off, especially since the cpu setting is suppose to be maxed.
Also the storage actively being used to pregen will delete about 10-20 gigs once the process stops I believe? I haven't touched it yet but its something I noticied but assumed was standard.
Could be because of the betternether mod?
I've never heard of that
I can confirm whatever process is happening after the pregen stopped, I hopped on the server and it was chugging processing pretty hard, even though it didn't say anything on the cpu usage end etc, so I stopped the server/
End of pregenning
and this is after Ive stopped and restarted server
Storage used dropped by like what is that 12-13 gigs?