Video walls are simple to implement using a frame buffer and simply using a section of the ram page as a source for each monitor. Digitise the video into a 2 port block of RAM, one block per colour. Then use some simple logic to read it out and replicate the data into another block of 2 port ram, one per monitor. You do the splitting of the image there and the data is then used to generate a genlocked composite video for the individual monitor. Each one just gets a video signal that has a magnified part of the original screen.
Lots of logic, and really needed to wait for LSI to make the programmable logic blocks fast enough to do the data splitting capable of running in real time. You always have a 3-4 frame delay on the video signal.
Just used a half rack filled with a fast video flash DAC ( available since the 1960's) and dual port ram ( early 1970's but expensive) and then a repeating section of video generators to make the wall signals from the frame buffer. The original ones from the 1960's that did frame rate conversion did not have a single IC inside, but used a large rack with thousands of Germanium transistors to build up the logic. they were still running in the UK until the switch off of 405 line TV a few years ago, though it was estimated only about 40 people still regularly watched the signal on a 405 line set. Not everything needs an IC to make it work.
Much easier now when you just send a serial data stream to each row of a large display and shift the data through at high speed in the frame blanking interval.