pub const fn vec128_b_index(m_pad: usize, token: usize, k_group: usize) -> usizeExpand description
2026-09-25: Offset of the scale for (token, k_group) in the B-operand
layout [L, m_pad], token index contiguous. The CUDA adapter
fp8_act_scale_to_kmajor (kernels/gb10/common/fp8_scale_transpose.cu)
writes the same index.