intel-analytics
Can Flashmoe run on an Intel shared GPU?
Hi @brownplayer , what is your OS and what is your target model ? Flashmoe only supports Linux for now.
That's really a pity. My system is Windows and I want to load qwen3-30b-3b. Do you have any plans to support Windows in the future?
---- Replied Message ---- | From | Ruonan @.> | | Date | 05/27/2025 10:12 | | To | intel/ipex-llm @.> | | Cc | Liu @.>, Mention @.> | | Subject | Re: [intel/ipex-llm] Can Flashmoe run on an Intel shared GPU? (Issue #13190) | rnwang04 left a comment (intel/ipex-llm#13190)
Hi @brownplayer , what is your OS and what is your target model ? Flashmoe only supports Linux for now.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
We may consider to support flashmoe for Windows later.
But for now, I guess you could run qwen3-30b-3b on Windows with ipex-llm llama.cpp (maybe pip install or portable zip), you may have a try with such cmd llama-cli -m xxx/qwen3-30b-3b -ngl 99 -ot exps=CPU -n 100 -c 2048.
Yes, I can now run qwen3-30b. However, FlashMoe can make it run partly in the video memory, which will be faster. Thank you for your contribution.
---- Replied Message ---- | From | Ruonan @.> | | Date | 05/27/2025 11:32 | | To | intel/ipex-llm @.> | | Cc | Liu @.>, Mention @.> | | Subject | Re: [intel/ipex-llm] Can Flashmoe run on an Intel shared GPU? (Issue #13190) | rnwang04 left a comment (intel/ipex-llm#13190)
We may consider support flashmoe for Windows later. But for now, I guess you could run qwen3-30b-3b on Windows with ipex-llm llama.cpp (maybe pip install or portable zip), you may have a try with such cmd llama-cli -m xxx/qwen3-30b-3b -ngl 99 -ot exps=CPU -n 100 -c 2048.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>