Conscious AI hypothetical

Suppose we have a conscious computer program; a program which, when executed, is conscious. We save the memory states and IO from the program at every stage of its execution, and then replay each state, one by one, on a second computer, without actually executing the program. Is the second computer conscious too?

In other words, if a computer is conscious, does its consciousness derive from the instructions it executes, or from the series of states it transitions between?

I feel that for strong AI to be possible at all, the machine has to be built up with a rigorous axiom of self preservation in its set of  operations from some level for the machine to ``validly'' become self-aware, a bit like adding conditional jump makes something Turing complete. Giving robots this axiom would be a  good idea for humans' lasting reign on earth.

And thats my shit theory, now someone give me DARPA funding, please!