May I ask what the purpose of a screen with 120 Hz is practical for? Isn't the eye's maximum only like 40 Hz?
I would like to know though... Does the server draw the camera angle for the player, or do they send the entire level to the client which then draws depending on your position? If its just the client, then you wouldn't need anything beyond a really good GPU.
As far as I know, the eye doesn't have a maximum perceivable frequency. In any case, if it did it would likely vary by person and light conditions etc.
In terms of my personal experience, games played in 120hz on my new acer monitor have massively reduced ghosting compared to my 6ms dell 24", games "feel" smoother and more fluid, but all these things hinge on personal perception.
For the 3d effect, a screen needs to be 3d so it can send out two 60hz frequencies, which the accompanying glasses then synchronise with to create the perception of depth.
In terms of the software I don't know how its done, it doesn't necessarily have to be coded by the developer, as Tribes:v (an older game) works in it for example. Nevertheless, developers can improve the implementation.
And yes, it does nom on your gpu quite heavily in 3d mode or 120hz mode.