pub fn q8_1_scratch_bytes(m: u32, k: u32) -> usizeExpand description
2026-09-25: q8_1 activation scratch bytes for [m, k]: 4 bytes per value with
k rounded up to 256, plus 1 MiB. ffn_act_q8 in
gpu-runtime/src/buffers/sizes.rs uses the same expression with k the
larger of the hidden and intermediate sizes.