[BugFix] Enhance test_pos_encoding to support execution on multi-devices

[wchen61]

"The test case in test_pos_encoding.py frequently fails when run in a multi-device environment.
FAILED test_pos_encoding.py::test_rotary_embedding[cuda:1-0-dtype0-32-120-17-8192-5-False] - RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:1! (when checking argument for argument index in method wrapper_CUDA__inde...
The issue arises because the rope created in this case does not specify a device, so it is always created on device 0, even if a default device is set. This patch fixes the problem, allowing the test case to pass when running on multiple devices.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

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 either: Add ready label to the PR or enable auto-merge.

🚀

[Isotr0py]

Thanks for the fix! The test passed locally with 2 GPUs with this fix.