missing APIs vs std::simd

Nobody has claimed this yet.

## Desirable, needs implementation (vectors)

- [x] `reduce_sum` https://github.com/linebender/fearless_simd/pull/357
- [x] `reduce_max`/`reduce_min` https://github.com/linebender/fearless_simd/pull/358

## Desirable, needs implementation (masks only)

- [x] Rotations https://github.com/linebender/fearless_simd/pull/360
- [ ] padded shifts

## Not investigated (TODO: move into another category)

- [ ] Horizontal bitwise reductions: `reduce_or`, `reduce_xor`, etc. #346
- [x] `reduce_product` https://github.com/linebender/fearless_simd/pull/361
- [ ] Signed int `saturating_abs`, `saturating_neg`, `is_positive`, `is_negative`, `signum`.
- [ ] `leading_zeros`, `trailing_zeros`, `leading_ones`, `trailing_ones` ([expensive emulation](https://chatgpt.com/share/6a9093b5-fe34-83eb-97d0-542ab698ee2b) but maybe still faster than scalar?)
- [ ] interleave and deinterleave on masks
- [ ] `to_degrees`, `to_radians` on floats
- [ ] Sign, NaN, infinity, finite, subnormal, and normal classification on floats
- [ ] `simd_clamp` on floats
- [ ] Ties-away-from-zero `round` on floats (round-ties-even is supported; ties away from zero has [poor hardware support](https://chatgpt.com/share/6a909862-5374-83ed-8b58-d263f5d23829))

## Compatibilty/convenience wrappers

Not sure if we should add all of these, the API surface is large enough as it is.

- [ ] `to_bits`/`from_bits` on floats (as an API compatibility wrapper around `bitcast`) https://github.com/linebender/fearless_simd/pull/370
- [x] `reverse` https://github.com/linebender/fearless_simd/pull/356
- [ ] `recip`
- [ ] `first_set`/`last_set` on masks (wrapper around `to_bitmask()` + scalar popcnt?)
- [ ] Boolean-scalar `&`, `|`, and `^` in either operand order on masks

## Tricky API questions

- [ ] `extract`
- [ ] `simd_swizzle!`
- [ ] generic `cast` between arbitrary same-lane-count vectors (we have explicit to_float/to_int and widen/narrow but not fully generic cast)
- [ ] lane-wise `abs_diff` that always needs to return unsigned values, and std doesn't implement for floats

## Not planned

- integer `div`, `rem`: no hardware support
- Masked load/store: `load_or_default`, `load_or`, `load_select` (only supported on AVX-512)
- Scatter/gather (impractically slow, often slower than scalar) #171
- Trigonometry, exponents, etc. (likely needs a separate crate with configurable trade-offs, not part of base fearless_simd)
- floating-point remainder: no hardware support, approximations available for special cases, [full support complex and expensive](https://chatgpt.com/share/6a9097ff-03d8-83eb-a15c-cc2f9d8493bc) (might bundle with trigonometry crate?)

Start by comparing the unchecked items in this checklist with the corresponding std::simd APIs, while noting the completed entries and linked pull requests. Choose one narrowly scoped missing API, then determine its implementation and verification points; done means the selected API is implemented and its behavior matches the intended std::simd compatibility.

Written by the indexing model from the issue text.