The engineering answer is to test and measure the actual hardware that is being considered because implementations often vary from theory in significant ways due to the quality of components, the competence of the design team, the design priority of energy saving, the design budget, the generation of technology, and other factors that abide in meat-space.

I mean the display panel A might be 20% more efficient than display panel B, but connected to a power supply that is half as efficient as B's.

Good luck.