Here's an old post I made on the subject,
Well, it's hardcoded and out there for public scrutiny :-p
If you're right about the 30 upgt cap, it might relate just to the player's movement (forward and sideways), and not the map axes.
However, it applies only to the x and y axes separately -not to their vector sum, which, per-se, is uncapped. However, its contituents momx/momy are not.
If you perform a "vanilla" wallrun accross a perfectly N/S or W/E wall, then only one of the axes will be activated significantly, hence the 30 mu/tic cap I measured. If there's a slope, then both axes can be activated, with a maximum possible vector sum of ca. 42.2 mu/tic.
This is also the reason why -among others- Doomguy can't reach the "nominal" impulse speed of 50 mu/tic.
I think I addressed those points in my edit. The individual momx/momy components are still capped at 30 mu/tics, and while their vector sum can reach 42.2 mu/tics, in theory the places where this would be possible should be few and far between (executing a perfect wallrun on a wall with a 45 degree y/x slope). Wallrunning strictly on one of the x or y axis would result in the single component 30 mu/tic cap.
In an unmodified port implementing the vanilla mechanics, yes I did.
Also, I measured actual distance between tics rather than momentums, but only on a two dimensional plane.
My measurements also came from distance (x-y) euclidean measurements but as I said, I didn't exactly push the wallrunning scenarios. In the few places in IWADs where I know it's possible (e.g. MAP01 corridor, E2M8 perimeter) it never went beyond 30 mu/tic.
Last edited by Maes on Nov 16 2012 at 19:36