This change updates the inline comments in SmithWatermanParVect.x10 to better describe the actual roles of the vector registers and arrays used in the Rognes-style Smith-Waterman inner loop. The code path itself does not change, but the documentation now explains concepts like H/E/F vector updates, gap penalty handling, and how state is carried between iterations, which makes the SIMD-style logic much easier to follow and maintain. Practical effect: no runtime behavior change, but future debugging and algorithm work should be less error-prone.