You keep assuming the power per transistor is the same for multi-state as binary, and we keep telling you that's probably not going to hold.On the other side, if you have 1 transistor doing the work of 2, it produces half the heat.
I think the point was about how/when intermediate states (i.e. neither the lowest nor highest) are quantized. In 2-state logic, you can easily reduce noise through saturation. It multi-state logic, it seems to me like it behaves closer to an analog circuit.Most of those difficulties come from analog computers not being digital.
Multivalued transistors are digital. They are not equivalent to analog computers.
I believe the way multi-bit NAND works is by using essentially using the equivalent of a DAC to write the cells and an ADC to read them. I wonder if you wouldn't need to periodically insert stages like that, to restore signal integrity, in large circuits using multi-state logic.
Last edited: