-
Notifications
You must be signed in to change notification settings - Fork 0
Troubleshooting
This is oversubscription. Vanilla's own worldgen/IO pool still runs alongside this mod's
dedicated chunk-builder pool — they don't share threads. If you set threads too high, both
pools can end up fighting for the same physical cores at once (e.g. flying through new terrain
triggers heavy worldgen and heavy chunk building simultaneously).
Fix: lower /multicoremagic threads <n> back toward the default (CPU cores − 1), or slightly
below it if you run other CPU-heavy background software (streaming/recording software, Discord
screen share, etc.) at the same time.
This is expected — the dedicated pool is wired in at the point the game (re)builds its chunk
renderer, which happens on world join or a resource/chunk reload (F3+A),
not instantly. /multicoremagic threads <n> is different: it resizes the pool that's already
running, so it takes effect immediately.
That means chunks are being requested faster than the pool can build them. Try raising threads
a bit (see the oversubscription note above for the ceiling), or check whether your view distance
is set unusually high for your CPU.
Older builds shipped without an icon (fixed in later releases). Update to the latest release from the Releases page if you still see this.
Please open an issue with your
Minecraft version, Fabric Loader version, and (if possible) the relevant lines from your log —
.minecraft/logs/latest.log.