Daniel,
That setup should work fine. And you are correct that the one DAQ device per one SCXI module in parallel rule does not apply to M Series devices with two connectors. One last thing I'd like to point out is that when you are determining your sample rate, you do not need to divide the total sample rate by 16, but only 8. Even when you are in differential mode (which you will be in this setup), which means of course that you are using two analog input channels for each measurement, it does not add the overhead of sampling two channels. So your max sampling rate would be more like 1MsS/8, or about 125kS/sec. Hope this info helps!
-Alan A.