Add device as parameter to TP and rotary_embedding functions

[chunyuan-w]

This PR adds device as parameter to TP and rotary_embedding functions.
For TP functions including get_tensor_model_parallel_world_size, get_tensor_model_parallel_rank and get_tp_group, we add device as parameter to distinguish the behavior between CPU and GPU.
For rotary_embedding, device="cuda" has been hard-coded. We add a device parameter and set the device of tensors accordingly.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[hmellor]

@chunyuan-w is this PR still relevant? If not, can we close it? (I'm just doing some housekeeping)

[chunyuan-w]

It's not relevant anymore