{
  "id": 7867461,
  "title": "Ollama Not Using GPU? Fix It on Linux, Windows and WSL",
  "url": "https://urgent.news/2026/09/16/ollama-not-using-gpu-fix-it-on-linux-windows-and-wsl",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-09-16T20:03:39.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/mrsaynothing/ollama-not-using-gpu-fix-it-on-linux-windows-and-wsl-3jb8"
  },
  "original_language": "en",
  "account": "To determine if Ollama is utilizing your GPU on Linux, Windows, or WSL2, follow these steps:\n\n1. Load a model in Ollama and check the output:\n- In one terminal, run `ollama run llama3.2 hello`.\n- In another terminal, execute `ollama ps`.\n\n2. Look at the \"PROCESSOR\" column in the `ollama ps` output. There are three possible states:\n- 100% GPU: The GPU is functioning correctly, and the model has successfully offloaded all layers.\n- 48%/52% CPU/GPU: The GPU is partially utilized, but the model or context size exceeds the available VRAM. Consider reducing the context size or using a quantized model.\n- 100% CPU: Inference is being processed by the CPU. The GPU either wasn't detected or was explicitly disabled.\n\n3. To identify the actual GPU Ollama is utilizing, examine the server log:\n- Run `journalctl -u ollama --no-pager | grep -i inference compute` on Linux.\n- On Windows, check the Ollama supported-GPU list to ensure your GPU is recognized.\n\n4. Common issues preventing GPU usage:\n- Outdated or incompatible NVIDIA driver: Update the driver, reboot, and restart the Ollama service.\n- Missing AMD group memberships (render and video) on Linux: Add the user to the appropriate groups and restart the Ollama service.\n- Unsupported AMD GPU: Ensure your AMD GPU is supported or attempt to override the detection using `HSA_OVERRIDE_GFX_VERSION`.\n- Pinned OLLAMA_LLM_LIBRARY: Remove the environment variable if it's causing issues with GPU detection.\n\n5. After applying driver or group fixes, restart the Ollama service to ensure it re-detects the GPU.\n\n6. If updating Ollama, the backend library may be pinned. Remove `OLLAMA_LLM_LIBRARY` if it's causing GPU usage issues.\n\n7. Ollama can indeed work in WSL2, provided the correct NVIDIA driver is installed and accessible from WSL2.",
  "summary": "TL;DR Run ollama ps while a model is loaded: the PROCESSOR column tells you the truth. 100% GPU means the GPU is fine and you can stop reading. A split like 40%/60% CPU/GPU means the model didn't fit in VRAM — use a smaller quant. 100% CPU means Ollama found no usable GPU: usually an outdated driver, a missing group membership (AMD on Linux), a pinned OLLAMA_LLM_LIBRARY , or a container started…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}