<algorithm>: Vectorize reverse_copy()

[StephanTLavavej]

Currently, we vectorize reverse():

STL/stl/inc/xutility

Lines 4599 to 4623 in 957fe99

	template <class _BidIt>
	void reverse(const _BidIt _First, const _BidIt _Last) { // reverse elements in [_First, _Last)
	_Adl_verify_range(_First, _Last);
	auto _UFirst = _Get_unwrapped(_First);
	auto _ULast = _Get_unwrapped(_Last);
	#if (defined(_M_IX86) || defined(_M_X64)) && !defined(_M_CEE_PURE) && !defined(_M_HYBRID)
	using _Elem = remove_pointer_t<decltype(_UFirst)>;
	constexpr bool _Allow_vectorization =
	conjunction_v<is_pointer<decltype(_UFirst)>, _Is_trivially_swappable<_Elem>, negation<is_volatile<_Elem>>>;
	if constexpr (_Allow_vectorization && sizeof(_Elem) == 1) {
	__std_reverse_trivially_swappable_1(_UFirst, _ULast);
	} else if constexpr (_Allow_vectorization && sizeof(_Elem) == 2) {
	__std_reverse_trivially_swappable_2(_UFirst, _ULast);
	} else if constexpr (_Allow_vectorization && sizeof(_Elem) == 4) {
	__std_reverse_trivially_swappable_4(_UFirst, _ULast);
	} else if constexpr (_Allow_vectorization && sizeof(_Elem) == 8) {
	__std_reverse_trivially_swappable_8(_UFirst, _ULast);
	} else
	#endif // (defined(_M_IX86) || defined(_M_X64)) && !defined(_M_CEE_PURE) && !defined(_M_HYBRID)
	{
	for (; _UFirst != _ULast && _UFirst != --_ULast; ++_UFirst) {
	_STD iter_swap(_UFirst, _ULast);
	}
	}
	}

We should extend this significant optimization to reverse_copy():

STL/stl/inc/algorithm

Lines 1975 to 1988 in 957fe99

	template <class _BidIt, class _OutIt>
	_OutIt reverse_copy(_BidIt _First, _BidIt _Last, _OutIt _Dest) {
	// copy reversing elements in [_First, _Last)
	_Adl_verify_range(_First, _Last);
	const auto _UFirst = _Get_unwrapped(_First);
	auto _ULast = _Get_unwrapped(_Last);
	auto _UDest = _Get_unwrapped_n(_Dest, _Idl_distance<_BidIt>(_UFirst, _ULast));
	for (; _UFirst != _ULast; ++_UDest) {
	*_UDest = *--_ULast;
	}
	_Seek_wrapped(_Dest, _UDest);
	return _Dest;
	}