RTX 3090 in Razer Core X Randomly Crashes on Ubuntu 22.04

Hey everyone, I’m having pretty severe issues with my RTX 3090 eGPU setup and am at my wits end.

My setup:

  • Framework Laptop 13 w/ Ryzen 7640u (BIOS version 03.05)
  • 96GB DDR5-RAM
  • Ubuntu 22.04 (Wayland) and 6.5.0-1025-oem kernel
  • Nvidia RTX 3090 in Razer Core X eGPU enclosure

The issue :

My eGPU regularly crashes (mostly under high load) forcing me to restart my laptop. Here’s what I’ve observed:

  1. The GPU mostly crashes during high load, for instance when running Stable Diffusion or an LLM. This can happen even after just 2 minutes of use.
  2. Other times, the 3090 runs fine for 20+ minutes at almost 100% load without issues.
  3. Occasionally, the GPU crashes even when there’s barely any load.
  4. When the crash occurs my whole system often becomes unresponsive as well.

What I’ve tried so far:

  1. Switch to open-kernel Nvidia drivers
  2. Limit the clockspeed of the 3090 with nvidia-smi -lgc 300,1750
  3. Blacklist the Nouveau drivers
  4. Add pcie_aspm=off to my kernel parameters.

I’ve attached two log files:

  1. https://pastebin.com/miKexVUJ: Contains journalctl, dmseg and other logs collected when the eGPU was working properly.
  2. https://pastebin.com/c8i1wiyW: Contains the same logs collected immediately after the eGPU crash.

I’m at my wits end and will have to sell the Razer Core X and the 3090 if I can’t fix this, but I’d much rather keep them, so I would be incredibly grateful if someone could help me :(

@generix I was hoping to hear what you think, since you were able to help others with their eGPU issues.