Improve performance of Series.to_numpy/to_cupy

[mroeschke]

Description

xref #11648

Essentially refactors Frame._to_array to short circuit some checks for a Frame with 1 column or ndim == 1

In [1]: import cudf

In [2]: s = cudf.Series(range(10000))

In [3]: %timeit s.to_cupy()
252 µs ± 3.47 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)  # PR

419 µs ± 2.21 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)  # branch 24.06

I needed to add Frame.ndim which will raise a NotImplementedError (until Frame actually becomes an ABC)

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[bdice]

Just one small fix.

[galipremsagar]

/merge