Ollama GPU usage validation

Home > Local system based AI tools > Ollama > Ollama GPU usage validation

Ollama GPU usage validation

1. Verify `/etc/systemd/system/ollama.service` has below line:

   Environment="OLLAMA_FLASH_ATTENTION=1"

2. Run below in one terminal to monitor GPU usage including processes using GPU

   watch nvidia-smi

3. Run some model on second terminal such as:

   ollama run deepseek-r1:8b

4. In nvidia-settings graphical output, click on GPU 0 - <GPU name> and see information about 'GPU Utilization:'
5. In third terminal if you use “ollama ps” we can see how much of CPU is being used and how much GPU

   ollama ps
   NAME              ID              SIZE      PROCESSOR          UNTIL
   deepseek-r1:8b    28f8fd6cdc67    6.3 GB    35%/65% CPU/GPU    4 minutes from now

Home > Local system based AI tools > Ollama > Ollama GPU usage validation