Add TurboQuant3/4 modes to quantized_scaled_dot_product_attention

[dedalien]

Integrates TurboQuant (arXiv 2504.19874) into the generic quant-SDPA infrastructure rather than as a standalone kernel.

New modes "turbo3" (3-bit) and "turbo4" (4-bit) use Lloyd-Max codebooks for the N(0,1) distribution of WHT-rotated, norm-normalised key coordinates. Codebooks are compile-time Metal constants; no runtime buffer needed.

Memory layout

• K/V: uint32-packed bit indices, same PackReader<N> path as affine
• k_scales/v_scales: per-vector float16 norms / √D
• group_size == head_dim (one scale per full D-element vector)
• Queries pre-rotated by caller via WHT

Changes

• primitives.h/cpp: TurboQuant3/TurboQuant4 in QuantizationMode enum, string parsing, quantization_mode_to_string explicit cases
• fast.cpp: TurboQuant validation branch (no biases, float scales, group_size == head_dim, head_dim ∈ {64,128,256}, GPU-only fallback)
• quantized_utils.h: QuantMode enum extension, QuantConfig/Dequant specialisations, Lloyd-Max codebook constants
• sdpa_vector.h: static_assert extended for bits=3; QUANT_SDPA_DISPATCH entries for D ∈ {64,128,256} gated on if constexpr (D >= group_size)
• scaled_dot_product_attention.cpp: quant_mode_to_int (TurboQuant3=4, TurboQuant4=5)
• python/src/fast.cpp: turbo3/turbo4 in mode docstring

Tests
12 sub-cases: D ∈ {64,128,256} × bits ∈ {3,4} × B ∈ {1,2}, each targeting a distinct Metal template instantiation. Explicit bfloat16, error path coverage.

Implementation notes

This implements TurboQuant_mse (Algorithm 1). TurboQuant_prod (Algorithm 2,
which adds a 1-bit QJL residual correction on the MSE residual) is not
implemented; at 3–4 bits the inner product bias of TurboQuant_mse is
empirically negligible (paper Section 4.1, Fig. 1: bias converges to zero
above 2 bits). This matches existing community implementations (arozanov,
rachittshah, sharpner).

The random rotation uses a Walsh-Hadamard Transform (WHT) rather than the
full random Gaussian rotation from the paper. WHT runs in O(D log D) vs
O(D²), is fully vectorizable on GPU, and gives equivalent practical quality
on real LLM key vectors.

Tested on Qwen3.6-27B (head_dim=256, 24Q/4KV, GQA=6), 24 GB unified memory Mac. Enables longer context generation on memory-constrained hardware by compressing the KV cache ~5×.

Note: This PR builds on ml-explore#3026.