pub fn eval_fused_block(fb: &FusedBlock, acts: &[i64], m: usize) -> Vec<i64>Expand description
Evaluate the fused block: O[i][c] = ReLU( ((A·W)[i][c] · M) >> S + bias[c] ) — one streamed pass, no
intermediate tensor written back. Byte-identical to computing the four stages separately (proven by the
lock), so the fusion is correctness-preserving; the win is that nothing lands in DRAM between stages.