My original point stands, also per TFA - performance gains from SMT are questionable for certain workloads. Whether OpenBSD prioritises absolute performance is besides the point - they benchmark against their own goals, not someone else's achievements.