This change is documentation-focused: it annotates the parallel and vectorized X10 implementations with clearer comments around matrix initialization, anti-diagonal scheduling, vector register flow, gap-penalty handling, and backtracking-related state. The underlying algorithms are largely unchanged, but the code is now much easier to read and maintain, especially in the more complex SIMD-style path where temporary vectors and shifts were previously hard to reason about. Practical effect: future contributors should have a much easier time understanding and modifying the parallel alignment code.