# GMKTec EVO-X2

128GB RAM + 2TB SSD + Ryzen AI MAX+ 395

## Neue Firmware

[https://strixhalo.wiki/Hardware/Boards/Sixunited\_AXB35/Firmware](https://strixhalo.wiki/Hardware/Boards/Sixunited_AXB35/Firmware)

BIOS: **1.11 20251017** [AXB35-02\_GMK\_SW1.11\_20251017.zip](https://strixhalo.wiki/Hardware/Boards/Sixunited_AXB35/Firmware/AXB35-02_GMK_SW1.11_20251017.zip)  
EC Firmware: **1.08 [EC-AXB35-02-1.08.zip](https://strixhalo.wiki/Hardware/Boards/Sixunited_AXB35/Firmware/EC-AXB35-02-1.08.zip)**

## Linksammlung

[https://www.reddit.com/r/MiniPCs/comments/1mptl53/gmktec\_evox2\_amd\_ryzen\_ai\_max\_395\_minipc\_with/](https://www.reddit.com/r/MiniPCs/comments/1mptl53/gmktec_evox2_amd_ryzen_ai_max_395_minipc_with/)

[https://www.reddit.com/r/LocalLLaMA/comments/1le951x/gmk\_x2amd\_max\_395\_w128gb\_first\_impressions/](https://www.reddit.com/r/LocalLLaMA/comments/1le951x/gmk_x2amd_max_395_w128gb_first_impressions/)

[https://www.reddit.com/r/GMKtec/comments/1ldtnbl/new\_firmware\_for\_evox2\_bios\_105\_ec\_106/](https://www.reddit.com/r/GMKtec/comments/1ldtnbl/new_firmware_for_evox2_bios_105_ec_106/)

[https://github.com/ollama/ollama/issues/12342](https://github.com/ollama/ollama/issues/12342) <

[https://github.com/ollama/ollama/issues/12062](https://github.com/ollama/ollama/issues/12062)

[https://www.reddit.com/r/ollama/comments/1nt5fcr/how\_do\_i\_get\_ollama\_to\_use\_the\_igpu\_on\_the\_amd\_ai/](https://www.reddit.com/r/ollama/comments/1nt5fcr/how_do_i_get_ollama_to_use_the_igpu_on_the_amd_ai/)

## AMD-Treiber installieren

Mit folgendem Tipp auf   
[https://www.reddit.com/r/ollama/comments/1nt5fcr/how\_do\_i\_get\_ollama\_to\_use\_the\_igpu\_on\_the\_amd\_ai/](https://www.reddit.com/r/ollama/comments/1nt5fcr/how_do_i_get_ollama_to_use_the_igpu_on_the_amd_ai/)   
auf folgenden Kommentar gestoßen  
[https://www.reddit.com/r/ollama/comments/1nt5fcr/comment/ngshfbr/](https://www.reddit.com/r/ollama/comments/1nt5fcr/comment/ngshfbr/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button)

> joseph\_sellers  
> •  
> 2mo ago  
> I am on a different stack (but same device). I am on Ubuntu 25.04, and using Docker, but I have Ollama working great.
> 
> I found an AMD APU specific fork of Ollama: [https://github.com/rjmalagon/ollama-linux-amd-apu](https://github.com/rjmalagon/ollama-linux-amd-apu)
> 
> On that fork, there is an open issue where someone says they have figured out how to get Ollama working on the exact device I am using: [https://github.com/rjmalagon/ollama-linux-amd-apu/issues/24](https://github.com/rjmalagon/ollama-linux-amd-apu/issues/24)
> 
> They have created a fork, of the fork, with the fix they found: [https://github.com/phueper/ollama-linux-amd-apu](https://github.com/phueper/ollama-linux-amd-apu)
> 
> Hopefully there are some clues there to help you get up and running on your setup.

dann auf [https://github.com/rjmalagon/ollama-linux-amd-apu?tab=readme-ov-file#check-amount-of-gtt-memory](https://github.com/rjmalagon/ollama-linux-amd-apu?tab=readme-ov-file#check-amount-of-gtt-memory) gekommen und folgendes cmd überprüft:

```
sudo dmesg | grep "amdgpu*"
```

danach ist mir aufgefallen, dass der init der amdgpu nicht geklappt hat:

<p class="callout danger">amdgpu: Fatal error during GPU init</p>

[![image.png](https://wiki.folkerts.it/uploads/images/gallery/2025-12/scaled-1680-/9Zqimage.png)](https://wiki.folkerts.it/uploads/images/gallery/2025-12/9Zqimage.png)

also unter [https://www.amd.com/en/support/download/linux-drivers.html](https://www.amd.com/en/support/download/linux-drivers.html) aktuellste Treiber für mein Ubuntu 24.04.03 heruntergeladen

```
sudo -i
wget https://repo.radeon.com/amdgpu-install/30.20.1/ubuntu/noble/amdgpu-install_7.1.1.70101-1_all.deb
chmod +x ./amdgpu-install_7.1.1.70101-1_all.deb
apt install ./amdgpu-install_7.1.1.70101-1_all.deb
amdgpu-install
reboot
```

danach

```
sudo dmesg | grep "amdgpu.*memory"
```

[![image.png](https://wiki.folkerts.it/uploads/images/gallery/2025-12/scaled-1680-/Lbqimage.png)](https://wiki.folkerts.it/uploads/images/gallery/2025-12/Lbqimage.png)

## Richtiges Ollama Dockerimage wählen

Für AMD GPUs muss das RCOm Image verwendet werden:

[https://hub.docker.com/r/ollama/ollama/tags?name=rocm](https://hub.docker.com/r/ollama/ollama/tags?name=rocm)