Renzo, would love to have a more detailed and andvance wiki entry with more examples if you're up for it some day. You seem to have more facts than guesses, unlike the rest of us

Thanks for all the suggestions.

During all those days my observation has only consolidated:
My aiming is *better* with *vsync enabled*, this is true for ezquake under Linux; in the windows client I can play with independent physics on and with vsync disabled, and my aiming is no worse than vice versa.

Yesterday, I compiled the most recent subversion ezquake, but to my disappointment I noticed that "show vidlag" only displayed a permanent "? fps" on my screen. Checking the code I noticed that cl_vsync_lag_fix doesn't seem to do anything under Linux; any chance to give up that win32-only status? Another suggestion:
I think, it would be much better to omit any empty variable from the code and the resulting binaries - if something hasn't yet been implemented for Linux (only for win32, for instance), the corresponding variable should be completely omitted for the Linux code, this way people are not playing around with things that only fool them.

I had also played with pushlatency some time ago, but checking the code I found that the variable does absolutely NOTHING. The same seems to be true for cl_vsync_lag_fix if you use Linux. It would be better to omit these variables as long as they have not been implemented yet. Just my proposal.


sempron

Man, thanks for all the info - I pretty much solved my tearing problem now.

I turned vid_vsync off, have cl_physfps at 77 (instead of 75 due vsync), have my monitor at 60 hz (which is fine if I don't have vsync on), cl_independentPhysics 1, and cl_maxfps 154.