Alright, now it is clear if we select one rank at a time with chip select. Then the gain is routing in dual rank with slower access rate compared to to channel single rank ? As we need to route the same set of signals including data lines, address lines, control and configuration lines between memory controller and both chips in dual rank. This can be two die in one package, DDP. To be clear here, then both chips will be in series to the memory controller, meaning that the same set of signals will be further router to the second chip, right ? As chips can be selected one at a time, which is seen like an additional address bit by the memory controller.
To get the same storage as single channel dual rank, we need to have two channels with single rank in which each memory will have parallel connection to the memory controller, like channel A and channel B. in each channel there will be complete set of data lines, address lines, control lines, and configuration lines. Each set will be routed between memory controller and the memory chips. This way we can have faster access but the trouble will be in routing two set of signals.