Hardware reset when playing with mod enabled
Lassebq opened this issue ยท 4 comments
Describe the bug
Whenever this mod is enabled and the game is running on the AMD integrated graphics, playing any world becomes a ticking time bomb. Literally. The issue happens in both overworld and the nether, but for whatever reason in the nether it's more consistent (usually takes 1-10 minutes to get crash)
After playing for a while, monitors immediately turn black, fans spin up, and the PC stops responding to any presses of power and reset buttons. The only way to shut down is to turn off power supply.
After rebooting some of accessed by the game files get corrupted or just turn into empty files. And it so happened that level.dat and level.dat_old are emptied after reboot. I have lost an important world that I put my time and effort into to this fault.
I still couldn't get a precise reason why this happens other than it being related to a combination of OpenGL, AMD integrated GPU, and this mod.
- I am not able to reproduce this bug if the game runs on the AMD Radeon RX 580 in this same PC.
- I am not able to reproduce this bug if the mod isn't active and the game is running on the AMD integrated graphics.
- I am able to reproduce this bug on both LWJGL 2 and LWJGL 3 (LWJGL 3 is achieved through translation layer)
- I am not able to reproduce this bug on sodium in modern minecraft and forks of sodium (such as neonium) for 1.12.2 when the game runs on iGPU
- I am not able to reproduce this bug using Optifine without Nothirium when the game runs on iGPU.
Reproducibility
Please try to reproduce the issue with as few other mods as possible. Then check one of the following checkboxes according to your results. To check a checkbox replace the space between the square brackets with an x (like this [x]) or create the issue and check the checkbox by clicking on it.
- I reproduced this issue with as few other mods as possible installed.
- I am unable to reproduce this issue consistently.
To Reproduce
- Go to the nether, because it's more consistent there
- Do anything. Load chunks, visit nether fortress, move camera. The issue does not seem to happen if you idle.
Since this is heavily tied to hardware, you'd likely need to have the same, or at least similar hardware to be able to reproduce this issue.
Screenshots/Videos
2025-07-20-15-11-02-808-2.mp4
Versions
Nothirium: 0.4.7-beta
RenderLib: 1.4.5
Forge: 14.23.5.2860
Minecraft: 1.12.2
Other mods (necessary to reproduce the bug): None
Hardware
Motherboard: Asrock B550 Phantom Gaming 4
CPU: Ryzen 5 4600G
RAM: 2x16 GB DDR4 3600Mhz
GPU: AMD Radeon Graphics (4600G)
VRAM: 1 GB dedicated
Software
OS: Arch Linux
Kernel: 6.16.0-rc2
OpenGL driver: Mesa 25.1.6 radeonsi
Log File
Due to a hardware crash latest.log either turns into a 0-byte file or cuts prematurely. Here's the prematurely cut log of the instance which subsequently caused a hardware reset
I would heavily advise NOT using a integrated gpu with Nothirium. Either way, this is wack
If you could share event viewer logs from when this happened, thatd be handy too
Managed to reproduce this same hardware reset using Vintagium for 1.12.2. Similarly to Nothirium the issue is way more common in the Nether.
if you're concerned about it being iGPU, on AMD Ryzen G series it's powerful enough to be compared to entry-level dedicated GPUs, so I'm not too worried about the performance without shaders. But I am worried about this one bug in particular.
I feel like a good starting point would be to use apitrace or similar to capture OpenGL (on the GPU which isn't prone to crashing), then replay using iGPU and see if it runs into a hardware reset in replay like during gameplay. Then I could point to the specifics of crash in GL calls.
