This question is bound to come up more frequently with the advent of rebuttal phases in major conferences. In such a system, authors get the comments, and sometimes the preliminary accept / reject votes, a little in advance to allow them to write an answer to the reviews, which is then - at least in theory - taken into account for the final decision. In such a system, it's fairly obvious when a paper is dead, and waiting for the final verdict is a waste of time in that case.
I don't think there is already a generally-understood moral code on whether you need to wait for the final rejection before submitting again. Formally withdrawing, as JeffE proposes, is a possibility, but given that the paper is, for all practical purposes, dead, a withdrawal seems to only add overhead at this point. Personally, I suspect a busy PC chair will just ignore your mail at this point.
Hence, I would not consider it a double-submission to immediately resubmit the paper. Just think of the paper as rejected and carry on. Note that this advice only holds if it is glaringly obvious that the paper will be rejected. Don't interpret this as permission to speculate if the reviews could still go either way, especially if the conference only gives you the textual comments but not the verdicts. Further, should your paper miraculously be accepted in the first conference anyway, you need to immediately withdraw from the second review process.