How technically impossible is it to combine two dp1.2 4k60 outputs into a single 4k120 input?

Question

My laptop has two Thunderbolt 3 ports with Alpine Ridge controllers that support Displayport 1.2 and can output a maximum 4k60 output. Is it possible to combine those two outputs to drive a single 4k120 display? Maybe it's possible to split the output into two virtual displays with xrandr and use each output for one half of the total image? Or building a custom board that combines two 1920x2160 120hz signals?

Answer 1

Disclaimer: much of theorycrafting in here, based on my limited knowledge about the standard. Take details of this answer with (big) pinch of salt.

Everything is possible in engineering if you have unlimited resources and loose enough requirements ;) And that's the case here, it's theoretically possible but in no way reasonable.

From my experience, I would say there are two possible ways:

1. Set up device able to send 4k120 and receive 2x 4k60. Decode the two signals and encode them to single one. Would need some expensive equipment (DisplayPort grabber with two ports, at least medium power *PU depending of requirements for latency (still it's near-impossible to go below 50-70ms latency). It would also require some way to synchronize signals to avoid tearing/jittering. But it should be quite easy while sending two halves of screen, adding an synchronizing line of pixels in the beginning of signal is quite easy. In practice: It would be way easier, more comfortable and cheaper to simply sell your old laptop and buy new one, with capability to send 4k120.
2. Become a Vesa member to get standard for Display Port electrical signals (or reverse engineer it). Create custom board which is able to receive daisy-chain signal, find the synchronization signal in them (if you use single device it's quite easy, it should be enough to blink a screens few times as a negotiation), and combine two signals using doubled clock. In this approach you would get unnoticable delays (even "slow" electronics would still introduce less than few ms). In practice: It's ridiculously expensive, both in workload and currency. You would buy plenty of top-notch laptops for resources spent even on creation of clumsy prototype.