Two important techniques have emerged recently for achieving high performance in processors. One is speculative execution, where instructions or sequences of instructions are executed before all the information needed to commit the instruction has been nailed down. The other is simultaneous multithreading (SMT), where the processor can issue instructions from multiple threads (or processes), potentially in the same cycle.

Do these approaches capture different aspects of/opportunities for parallelism, or are they different ways at getting at the same results? If both were implemented in a single system, would you expect their projected improvement to be additive? Why or why not?