Add Neon implementations of std::reverse, std::reverse_copy

[hazzlim]

Benchmark numbers (relative performance compared to current baseline, > 1 is better)

name	MSVC Speedup	Clang Speedup
rstd::uint8_t/3449	12.17	14.00
rstd::uint8_t/63	4.35	5.69
rstd::uint8_t/31	2.22	2.85
rstd::uint8_t/15	1.24	1.48
rstd::uint8_t/7	0.59	0.82
rstd::uint16_t/3449	6.24	7.00
rstd::uint16_t/63	3.01	4.07
rstd::uint16_t/31	2.32	2.85
rstd::uint16_t/15	1.24	1.55
rstd::uint16_t/7	0.61	0.75
rstd::uint32_t/3449	2.47	2.18
rstd::uint32_t/63	2.58	3.20
rstd::uint32_t/31	1.98	2.41
rstd::uint32_t/15	1.22	1.48
rstd::uint32_t/7	0.6	0.85
rstd::uint64_t/3449	1.36	1.22
rstd::uint64_t/63	2.16	2.28
rstd::uint64_t/31	1.85	2.15
rstd::uint64_t/15	1.16	1.65
rstd::uint64_t/7	0.9	1.24
rcstd::uint8_t/3449	13.07	0.99
rcstd::uint8_t/63	3.99	1.27
rcstd::uint8_t/31	2.18	1.18
rcstd::uint8_t/15	1.15	0.97
rcstd::uint8_t/7	0.57	0.65
rcstd::uint16_t/3449	6.39	1.00
rcstd::uint16_t/63	3.99	1.41
rcstd::uint16_t/31	2.72	1.44
rcstd::uint16_t/15	1.61	1.38
rcstd::uint16_t/7	0.88	1.20
rcstd::uint32_t/3449	3.26	0.98
rcstd::uint32_t/63	3	1.34
rcstd::uint32_t/31	2.23	1.55
rcstd::uint32_t/15	1.65	1.83
rcstd::uint32_t/7	1.08	1.29
rcstd::uint64_t/3449	2.66	1.24
rcstd::uint64_t/63	2.22	1.31
rcstd::uint64_t/31	1.81	1.27
rcstd::uint64_t/15	1.36	1.29
rcstd::uint64_t/7	1	1.41

[AlexGuteniev]

Benchmark numbers

clang-cl appears to auto-vectorize reverse_copy, but not reverse

[StephanTLavavej]

I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed.

[StephanTLavavej]

😻 !uoy knahT

[hazzlim]

Thanks for reviewing @AlexGuteniev and @StephanTLavavej 🐈‍⬛