Efficient LLM inference solution on Intel GPU

Re-assign community