docs: clarify cuda async execution model

[0ax1]

Summary

Besides adding docs for the execution model, this includes notes on pageable and pinned physical memory. Further, move_to_device is dropped in favor of copy_to_device.

[codspeed-hq]

Merging this PR will improve performance by 10.94%

⚡ 1 improved benchmark
✅ 985 untouched benchmarks
⏩ 1426 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	map_each[BufferMut<i32>, 128]	887.2 ns	799.7 ns	+10.94%

---

_{Comparing ad/cuda-async-docs (64e9d97) with develop (ab55ce4)}

Footnotes

1. 1426 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩