feat: turboquant encoding for vectors #7167

Performance Gate Passed

✅ 1106 untouched benchmarks
🆕 20 new benchmarks
⏩ 1522 skipped benchmarks¹

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
🆕	Simulation	`turboquant_compress_dim1024_2bit_mse`	N/A	66.7 ms	N/A
🆕	Simulation	`turboquant_compress_dim1024_2bit_qjl`	N/A	136.4 ms	N/A
🆕	Simulation	`turboquant_compress_dim1024_8bit_mse`	N/A	730.6 ms	N/A
🆕	Simulation	`turboquant_compress_dim128_4bit_mse`	N/A	13.1 ms	N/A
🆕	Simulation	`turboquant_compress_dim1024_4bit_mse`	N/A	103.7 ms	N/A
🆕	Simulation	`turboquant_compress_dim768_4bit_qjl`	N/A	160.7 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_2bit_mse`	N/A	38 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_8bit_qjl`	N/A	85.8 ms	N/A
🆕	Simulation	`turboquant_compress_dim768_4bit_mse`	N/A	102.8 ms	N/A
🆕	Simulation	`turboquant_compress_dim128_4bit_qjl`	N/A	20.5 ms	N/A
🆕	Simulation	`turboquant_decompress_dim768_4bit_mse`	N/A	37.4 ms	N/A
🆕	Simulation	`turboquant_compress_dim1024_4bit_qjl`	N/A	162 ms	N/A
🆕	Simulation	`turboquant_compress_dim1024_8bit_qjl`	N/A	481.7 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_8bit_mse`	N/A	38 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_2bit_qjl`	N/A	85.9 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_4bit_mse`	N/A	38 ms	N/A
🆕	Simulation	`turboquant_decompress_dim1024_4bit_qjl`	N/A	85.9 ms	N/A
🆕	Simulation	`turboquant_decompress_dim768_4bit_qjl`	N/A	80.9 ms	N/A
🆕	Simulation	`turboquant_decompress_dim128_4bit_mse`	N/A	5 ms	N/A
🆕	Simulation	`turboquant_decompress_dim128_4bit_qjl`	N/A	11.1 ms	N/A

_{Comparing claude/admiring-lichterman (c09177b) with develop (e8cd130)}

1522 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩