Add reduce_scatter CCL example (27)

[Copilot]

Adds examples/27_ccl_reduce_scatter/ alongside the existing CCL examples (all_reduce, all_gather, all_to_all).

Changes

• examples/27_ccl_reduce_scatter/example.py — runnable example matching the structure of examples 24–26. Each rank contributes an (M, N) tensor filled with (rank + 1); ctx.ccl.reduce_scatter writes the reduced result only to each rank's assigned tile partition.

CCL reduce scatter is already covered by test_reduce_scatter_with_groups in tests/ccl/test_process_groups.py, so no new test file is added.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Add reduce scatter example</issue_title>
<issue_description>Add reduce scatter example similar to examples 24, 25 and 26</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Add reduce scatter example #438

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[mawad-amd]

@copilot we have a test for ccl reduce scatter already can you confirm

[Copilot]

@copilot we have a test for ccl reduce scatter already can you confirm

Yes — the base branch already had test_reduce_scatter_with_groups in tests/ccl/test_process_groups.py (covering process-group scoping). This PR additionally adds a dedicated tests/ccl/test_reduce_scatter.py with broader parametrization over dtypes, shapes, and both tile-distribution modes (striding and block), consistent with the existing test_all_reduce.py / test_all_gather.py test files.

[mawad-amd]

@copilot remove this file if we already have the tests in plac

[Copilot]

Done — removed tests/ccl/test_reduce_scatter.py in fe1a915.