Upgraded to xsimd 14

[JohanMabille]

No description provided.

[codspeed-hq]

Merging this PR will not alter performance

⚡ 4 improved benchmarks
❌ 1 (👁 1) regressed benchmark
✅ 250 untouched benchmarks

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	math_xtensor_1[expm1_fn, xtensor<double, 2>][64]	126.1 µs	99.4 µs	+26.86%
⚡	Simulation	xshape_initializer[std\:\:array<std\:\:size_t, 4>]	180.8 ns	151.7 ns	+19.23%
⚡	Simulation	math_xtensor_1[tanh_fn, xtensor<double, 2>][64]	119.3 µs	104.2 µs	+14.56%
⚡	Simulation	math_xtensor_1[sinh_fn, xtensor<double, 2>][64]	113.8 µs	101.6 µs	+12.02%
👁	Simulation	xshape_access[std\:\:array<std\:\:size_t, 4>]	125.3 ns	154.4 ns	-18.88%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing JohanMabille:xsimd_14 (2121095) with master (1b5325b)}