[TPU] Add support for PJRT from PyTorch/XLA 2.0

[Liyang90]

What does this PR do?

Fixes #\17350

To use PJRT with XLAStrategy, simply set PJRT_DEVICE env var, e.g. PJRT_DEVICE=TPU.

To run on a TPU Pod with PJRT, xla_dist is not needed anymore. You can directly issue the same command to all TPU workers:

gcloud compute tpus tpu-vm ssh <TPU name> --zone=<TPU zone> --project=<project name> --worker=all --command="PJRT_DEVICE=TPU python3 <your script>"

For more details on PJRT, please visit: https://github.com/pytorch/xla/blob/r2.0/docs/pjrt.md

Before submitting

• Was this discussed/agreed via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

[Borda]

lets make also the TPU as experimental as it uses XLA experimental

[Liyang90]

@will-cromar for review