Use merge_qkv to replace fused_qkv for llama2

[rnwang04]

Description

1. Why the change?

• Use merge_qkv to replace fused_qkv for llama2 to improve batch performance

• update to use new rope

• validate Llama-2-7b-chat-hf when transformers 4.36.2 with IPEX_LLM_QUANTIZE_KV_CACHE=0

• validate Llama-2-7b-chat-hf when transformers 4.36.2 with IPEX_LLM_QUANTIZE_KV_CACHE=1

• validate Llama-2-7b-chat-hf when transformers 4.31.0 with IPEX_LLM_QUANTIZE_KV_CACHE=0

• validate Llama-2-7b-chat-hf when transformers 4.31.0 with IPEX_LLM_QUANTIZE_KV_CACHE=1

• validate Llama-2-7b-chat-hf when transformers 4.41.0 with IPEX_LLM_QUANTIZE_KV_CACHE=0

• validate Llama-2-7b-chat-hf when transformers 4.41.0 with IPEX_LLM_QUANTIZE_KV_CACHE=1

• validate Llama-2-7b-chat-hf when transformers 4.36.2 on CPU

2. User API changes

No change.

3. Summary of the change

• Use merge_qkv to replace fused_qkv for llama2 to improve batch performance
• update to use new rope
• Fix CPU ut

4. How to test?

• Unit test: Please manually trigger the PR Validation here by inputting the PR number (e.g., 1234). And paste your action link here once it has been successfully finished.

[rnwang04]

PR validation: https://github.com/intel-analytics/ipex-llm-workflow/actions/runs/10265377508/job/28401561843

[rnwang04]

PR validation is passed: https://github.com/intel-analytics/ipex-llm-workflow/actions/runs/10276986872