Constant W8A16_M32_K_BLOCK 

pub const W8A16_M32_K_BLOCK: u32 = 128;

2026-09-25: K granularity: one FP8 scale block per K-step (PM32_K_STEP == PM32_FP8_BLOCK), which is why K % 128 == 0 is required rather than padded.