That is an interesting question.
As noted both laptops have the same active and passive cooling system, comprised of the following: Two large cooling fans that exhaust out at both ends of the rear vent located along the screen hinge. A primary air intake at the center of the rear vent (between the two exhaust vents) and two additional sets of intake vents along side edges. At the exhaust of each fan a radiator fins transfers heat to the fan’s exhaust air. A single heat pipe then connects at each end to the two radiator fins. For the iGPU model the heat pipe is only connected to a single heat sink at the CPU/iGPU processor, and for the dGPU model the heat pipe also connects to an additional and separate heat sink at the dedicated GPU processor. Since the heat pipe is tasked with transferring heat away from the processors, then its performance under the two configuration could contribute to one being more efficient than the other.
The CPU/iGPU processor is located roughly in the middle of the heat pipe so for the single processor configuration the heat pipe splits the heat about evenly to each side. Just as long as the processing tasks are either CPU or GPU intensive then this configuration will benefit from being served by the two fans. However, if workload simultaneously tasks both the CPU and GPU then the heat of each processor will immediately impact the other and the single processor heat sink and heat pipe connection will likely become a bottle neck, forcing heat to pull up at the processors, which could lead to throttling.
By separating the two processors during CPU and GPU intensive workloads each heat sink and heat pipe connection will have less heat to transfer, plus each processor will essentially get a dedicated single fan for cooling, thus preventing heat pooling up and leading to throttling. With the dGPU setup the CPU will most likely generate the majority of heat and in that setup the presence of the dGPU could act as a heat barrier, limiting the CPU to just a single cooling fan, as opposed the the single processor configuration that would split cooling more evenly between the two fans. On the flip side, during graphics intensive workloads with the dGPU configure, the dGPU’s close proximity to the one cooling fan may contribute to efficient cooling with zero impact on the CPU.
In addition to the internal cooling components, the laptop’s entire aluminum enclosure acts as a heat sink, helping to cool the interior by transferring heat to the exterior surroundings. The two processors configurationwill be more effective at transferring and dissipating waste heat into the aluminum enclosure than the single processor configuration.
So overall the two process dGPU configuration likely makes better use of the cooling system. Of course that is only evaluating the heat removal side of the equation, the amount of potential heat generation also needs to be considered. The single CPU/iGPU processor has a slower speed so it will generate less heat and therefore be less likely to exceed the cooling system’s built-in cooling capacity. The separate CPU and dGPU processor models have much higher speeds, and while you cannot run both the iGPU and dGPU simultaneously, the potential heat generation is still much greater, and more likely to overwhelm the cooling system even though it may have a slightly more efficient use of the cooling system.
The most important question is whether you need the extra graphics processing power of the dGPU configuration and/or the extra central processing power of the higher-end CPU. If you do need the extra power then great, you might get a slightly more efficient configuration and you get more work done with the higher process speeds when throttling is not a factor. However, if you do not need that extra speed then you save money and the lower processor speeds will generate much less heat so the slightly less efficient configuration probably won’t even come into play.