→ latency optimizations: Two optimizations reduce wall‑clock time: parallelizing across devices and feeding the public key to the quantum computer mid‑run after a generic setup phase