Make use of GetMany to improve perf of CopyTo

[Manodasan Wignarajah (manodasanW)]

Measured IList<T>.CopyTo over 100,000-element native IVector<T> instances using the CollectionsPerf.GetMany* benchmarks.

Scenario	Before	After	Improvement
int	10.55 ns/item	0.111 ns/item	~95× faster
string	61.27 ns/item	37.99 ns/item	~38% faster
projected object	88.24 ns/item	80.49 ns/item	~9% faster

Measured a native component calling IVector<int>.GetMany on a managed int[] containing 100,000 elements.

Implementation	Mean per item	Relative
Baseline — IList<T> indexer loop	0.652 ns	1.00×
Optimized — direct span copy	0.116 ns	0.18×

The optimized path is approximately 5.6× faster, reducing per-item cost by ~82%, with no managed allocations reported.

Scenario	Baseline	GetMany optimized	Improvement
IList<int>.ToArray()	11.20–11.23 ns/item	0.46–0.49 ns/item	22.7–24.6× faster
IList<string>.ToArray()	104.2–104.7 ns/item	73.4–79.2 ns/item	24–30% faster
IList<WrappedClass>.ToArray()	190.3 ns/item	176.3–177.1 ns/item	7.0–7.4% faster
IVectorView<int>.ToArray()	11.16–11.26 ns/item	0.49–0.52 ns/item	21.5–22.9× faster
IVectorView<WrappedClass>.ToArray()	211.9–224.7 ns/item	177.7–191.5 ns/item	14.8–16.1% faster

ToArray() benefits because LINQ allocates the destination array and invokes ICollection<T>.CopyTo(), which now uses native GetMany.

[azure-pipelines]

Azure Pipelines:
There may be pipelines that require an authorized user to comment /azp run to run.