[Hardware][TPU]Enable ragged paged attention kernel and resolve recompilation issue

[yaochengji]

• enable ragged paged attention kernel
• resolve recompilation issue
• change the layout of kv cache from (num_kv_heads, num_blocks, block_size, head_size) to (num_blocks, block_size, num_kv_heads, head_size)

[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.

🚀

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @yaochengji.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[yaochengji]

This PR depends on pytorch/xla#8797, I will update the torch/xla nightly version tomorrow

[mgoin]

Should this shift depending on device?

[yaochengji]

Yes, @vanbasten23 is tuning the kernel parameter, will update it later.