Half-Life tweak

[ianshepherd]

Hi all,

Apologies if this has already been discussed, but I've been using an autoexec.cfg based on Kruzin's for ages, and just found that *disabling* gl_ztrick boosted my timerefresh scores by about 50%. This is with a G400 16 MB vanilla, no o/c, using the TGL on a Celeron 500. Play seems smoother too. It never used to make any difference ( before the TGL ). Anyone else find this ?

Cheers

Ian

PS. Seems the gl_texsort 1 tweak no longer has any effect...?

[Nitromind]

i made some timerefreshs....

1024x768x32

gl_ztrick 1
gl_texsort 1
i get 46 fps

gl_ztrick 1
gl_texsort 0 (default)
i get 68fps

gl_ztrick 0 (default)
gl_texsort 0 (default)
i get 77fps

so i would say with TGL there is no need for Kruzin's optimised cfg file

[ianshepherd]

Hmm,

well a few timedemo tests _don't show any difference ( which is how it should be, since this is a 3dfx variable, right ? ) but the timerefresh shows a clear difference - anyone have any ideas what's going on ?

Just curious...

Ian

[Kruzin]

I have not tested this too much with the TurboGL, but I can tell you that timerefresh is terribly unreliable as a form of benchmarking the game, as Ian's second post clearly shows...

[ianshepherd]

Hi Kruzin,

I agree with you _except the game does feel smoother in play - and timerefresh has to count for something, doesn't it ?
My guess is that the game feels smoother because all the "looking around" and sudden turns you make during play ( which is what a timerefresh emulates ? ) have a better framerate, even though in combat the fps stays similar...?

I find timedemo can be deceptive as well - my scores at 1024x768 are identical to the 800x600, but in play it's clearly less smooth. I guess because it's better to run with some overhead - maybe again this is where the gl_ztrick 0 helps ?
Do you see any difference with the TGL ?

Cheers

Ian