Add vLLM EP launch script and SGLang scripts for Intel irdma NICs

[manojgop]

• Add launch_vllm_intel_nic.sh for multi-node vLLM Expert Parallel with Intel irdma RoCE NICs (Qwen3-30B-A3B-FP8)
• Add common_env_intel_nic.sh with shared Intel NIC env config
• Add Qwen3-30B_uep_intel_nic.sh for SGLang EP on Intel NICs
• Set LOCAL_WORLD_SIZE to fix cross-node CUDA IPC handle error

Description

Please include a summary of the changes and the related issue.

Fixes # (issue)

Type of Change

• Bug fix
• New feature
• Documentation update

How Has This Been Tested?

Include any tests here.

• Unit tests
• Integration tests
• Manual testing

Checklist

• I have run format.sh to follow the style guidelines.
• I have run build.sh to verify compilation.
• I have removed redundant variables and comments.
• I have updated the documentation.
• I have added tests.

[MaoZiming]

@manojgop Thank you! Do you want to share the results or performance in this PR or README.md?

[manojgop]

@manojgop Thank you! Do you want to share the results or performance in this PR or README.md?

I'll raise one more PR for P-D disaggregation. Shall I update the the README.md with that PR ? This script is similar to existing script except that environment variables are Intel NIC specific. Don't want to share performance results now since this is tested with a smaller setup with only 1 GPU per node.

[MaoZiming]

@manojgop Thank you! Do you want to share the results or performance in this PR or README.md?

I'll raise one more PR for P-D disaggregation. Shall I update the the README.md with that PR ? This script is similar to existing script except that environment variables are Intel NIC specific. Don't want to share performance results now since this is tested with a smaller setup with only 1 GPU per node.

I see. sounds good.