Apparently Itanium works quite well for LLMs https://medium.com/@tglozar/running-llama-inference-on-intel-itanium-part-1-be62ff3f5c2f https://medium.com/@tgloza

Apparently Itanium works quite well for LLMs https://medium.com/@tglozar/running-llama-inference-on-intel-itanium-part-1-be62ff3f5c2f https://medium.com/@tglozar/running-llama-inference-on-intel... Which makes sense I suppose.