Why i cannot use my gpu

[jackphj]

i created conda env llm-cpp
and then
conda activate llm-cpp
init-ollama.bat
$env:no_proxy = "localhost,127.0.0.1"; $env:ZES_ENABLE_SYSMAN = "1"; $env:OLLAMA_NUM_GPU = "999" ;$env:SYCL_CACHE_PERSISTENT = "1"; $env:SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS = "1"
and then
.\ollama.exe serve

as we can see, it doesn't match the situation in the link

CPU: intel ultra7 258v
GPU: intergated Arc 140v

link:
https://github.com/intel-analytics/ipex-llm/blob/main/docs/mddocs/Quickstart/ollama_quickstart.md

[sgwhat]

Hi @jackphj,

May I ask why you think ollama isn't running on a GPU. Have you tested running a model with ollama? If you have more questions, please provide the complete log from ollama server side.

[jackphj]

Hi @jackphj,

May I ask why you think ollama isn't running on a GPU. Have you tested running a model with ollama? If you have more questions, please provide the complete log from ollama server side.请问您为什么认为 ollama 不在 GPU 上运行。您是否测试过使用 ollama 运行模型？如果您还有其他问题，请提供来自 ollama 服务器端的完整日志。

Because I found that the log in the picture above says OLLAMA_INTEL_GPU:FALSE.

[sgwhat]

Because I found that the log in the picture above says OLLAMA_INTEL_GPU:FALSE.

Sorry that's a confusing and useless log, it's meaningless, I will delete it later. You may run your model via ollama run <models> to check.