[FEA] Add all PDSH queries for cudf.pandas

[mroeschke]

cudf_polars currently contains all 22 PDSH queries in https://github.com/rapidsai/cudf/blob/branch-25.12/python/cudf_polars/cudf_polars/experimental/benchmarks/pdsh.py. We should also implement these queries to be run with cudf.pandas in order to:

1. Have some visibility of runtime performance over time, namely a proxy to detect unexpected fallback behaviors of "common" operations
2. A potential, consistent source from which we can socialize the performance benefits of cudf.pandas over pandas (e.g. documentation, blogs, etc.)

These benchmarks should be implemented in a manner that allows them to be run on a consistent (e.g. nightly) basis like the cudf_polars PDSH benchmarks.