Baike.dev
Log in
Back to tool/Back to issues
#1345·project-nomad

[Bug]: iGPU passthrough on AMD mini pc shows a warning "GPU Not Accessible to AI Assistant,"

Author: davidvermillionCreated Sep 14, 2026Updated Sep 14, 2026
Labelsbug

Issue Category

AI Assistant (Ollama)

Bug Description

iGPU passthrough on AMD mini pc shows a warning "GPU Not Accessible to AI Assistant," fix button doesn't seem to have any affect.

Steps to Reproduce

Steps:

  1. Go to settings
  2. Click System
  3. A warning is displayed: GPU Not Accessible to AI Assistant

Your system has an AMD GPU, but the AI Assistant can't access it. AI is running on CPU only, which is significantly slower.

  1. Click the button marked: Fix: Reinstall AI Assistant
  2. A notification will popup indicating the AI assistant is being reinstalled, then, the page will reload

Expected Behavior

Expect: iGPU support to be enabled and the warning should not be displayed.

Actual Behavior

Result: The error message remains

NOMAD Version

1.35.0-rc.1

Operating System

Ubuntu 26.04 (Resolute Raccoon)

Docker Version

No response

Do you have a dedicated GPU?

No

GPU Model (if applicable)

No response

System Specifications

Project NOMAD Debug Info

App Version: 1.35.0-rc.1 Environment: production

System: OS: Ubuntu 26.04.1 LTS Hostname: nomad Kernel: 7.0.0-31-generic Architecture: x64 Docker Engine: 29.8.0 Uptime: 21h 53m

Hardware: CPU: Ryzen 7 H 255 w/ Radeon 780M Graphics (16 cores) RAM: 59.6 GB total, 3.3 GB used, 56.4 GB available GPU: Device 1900 (512 MB VRAM) GPU Passthrough: passthrough_failed (amd) Disk: 937.8 GB, 24% used, NVME

Storage: Host storage root: /opt/project-nomad/storage Container path: /app/storage NOMAD_STORAGE_PATH: /opt/project-nomad/storage Kiwix library: 66 book(s)

Installed Services: Information Library (nomad_kiwix_server): running Education Platform (Gen 2) (nomad_kolibri_2): running AI Assistant (nomad_ollama): running

Internet Status: Online Update Available: No (1.35.0-rc.1 is latest)

Auto-Update: Core: Enabled Apps: Enabled Content: Enabled Browser: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:155.0) Gecko/20100101 Firefox/155.0

Service Status (if relevant)

No response

Relevant Logs

bash
produ@nomad:~$ docker logs nomad_ollama
time=2026-09-14T17:17:26.381Z level=INFO source=routes.go:1965 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: LLAMA_ARG_FIT: LLAMA_ARG_FIT_TARGET: NO_PROXY: OLLAMA_CONTEXT_LENGTH:8192 OLLAMA_DEBUG:INFO OLLAMA_DEBUG_LOG_REQUESTS:false OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:true OLLAMA_GO_TEMPLATE:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_IGPU_ENABLE:1 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE:q8_0 OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MAX_TRANSFER_STREAMS:4 OLLAMA_MODELS:/root/.ollama/models OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD:true OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:true ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
time=2026-09-14T17:17:26.381Z level=INFO source=routes.go:1967 msg="Ollama cloud disabled: true"
time=2026-09-14T17:17:26.381Z level=INFO source=images.go:957 msg="total blobs: 9"
time=2026-09-14T17:17:26.381Z level=INFO source=images.go:964 msg="total unused blobs removed: 0"
time=2026-09-14T17:17:26.382Z level=INFO source=routes.go:2022 msg="Listening on [::]:11434 (version 0.34.0)"
time=2026-09-14T17:17:26.382Z level=INFO source=model_recommendations.go:177 msg="model recommendations cache sleep scheduled" wait=3h48m57.6690491s consecutive_failures=0
time=2026-09-14T17:17:26.383Z level=INFO source=runner.go:60 msg="discovering available GPUs..."
time=2026-09-14T17:17:26.383Z level=INFO source=model_list_cache.go:112 msg="model list cache hydration complete" models=2 failures=0 elapsed=1.508322ms
time=2026-09-14T17:17:26.591Z level=WARN source=amd.go:470 msg="dropping ROCm device — no rocblas support for gfx target" device=ROCm0 gfx_target=gfx1103 supported="map[gfx1030:true gfx1100:true gfx1101:true gfx1102:true gfx1150:true gfx1151:true gfx1200:true gfx1201:true gfx908:true gfx90a:true gfx942:true gfx950:true]" hint="set HSA_OVERRIDE_GFX_VERSION to map to a supported target"
time=2026-09-14T17:17:26.591Z level=INFO source=types.go:50 msg="inference compute" id=cpu library=cpu compute="" name=cpu description=cpu libdirs=ollama driver="" pci_id="" type="" total="59.6 GiB" available="59.6 GiB"
time=2026-09-14T17:17:26.591Z level=INFO source=routes.go:2072 msg="vram-based default context" total_vram="0 B" default_num_ctx=4096
produ@nomad:~$

Browser Console Errors (if UI issue)

javascript

Screenshots

No response

Additional Context

docker logs nomad_qdrant.txt docker logs nomad_admin.txt

Pre-submission Checklist

  • I have searched for existing issues that might be related to this bug
  • I am running the latest version of Project NOMAD (or have noted my version above)
  • I have redacted any personal or sensitive information from logs and screenshots
  • This issue is NOT related to running NOMAD on an unsupported/non-Debian-based OS

Source: Crosstalk-Solutions/project-nomad

View original on GitHubView discussion on GitHub