multiple times but it is written in such an obtuse way it is really hard to get anything out of it. It does not help that if you skim engine sources you find bunch of configuration directives that seemingly either duplicate the existing behavior or control it the same way but in different place:
r.GTSyncType r.OneFrameThreadLag r.FinishCurrentFrame
What a mess. It used to be simple before, you set queued frames to 0 and you get earliest possible frame, the end.
From what I understood,
r.GTSyncType 2 only relevant for cases with vsync, value
1 would enable old behavior that people lovingly call “laggy ue mess”. So I’m assuming I would want value
0. Now how would the rest 2 contribute to that?
Man, it is hard to test latency without these sweet LDAT kits.