Falcosoft wrote:

derSammler wrote:

Any sources for that? I'm using a GF2 GTS with 71.84 and did not run into any issues so far.

https://www.youtube.com/watch?v=HRhm4aGNI3o

While I respect Phils' work, a single person testing such a thing on a single system does not backup the statement. The behaviour found may be caused by the mainboard chipset, the chipset drivers (which may not play nicely with newer nvidia drivers) or the CPU. Unless you do a broad test, that single result has no validity.

derSammler wrote:

Falcosoft wrote:

derSammler wrote:

Any sources for that? I'm using a GF2 GTS with 71.84 and did not run into any issues so far.

https://www.youtube.com/watch?v=HRhm4aGNI3o

While I respect Phils' work, a single person testing such a thing on a single system does not backup the statement. The behaviour found may be caused by the mainboard chipset, the chipset drivers (which may not play nicely with newer nvidia drivers) or the CPU. Unless you do a broad test, that single result has no validity.

You asked for 'any sources' and that's one. Even one of the most comprehensive tests. Leileilol claims he experienced the same, me too. That's 3. But universal statements cannot be verified by any finite observations so even if there would be 1000 the statement still were not fully backed up. So you should simply test this for yourself. I think it's enough to test e.g. with 30.82 or 45.23 and your current version. Then you can partly verify or fully falsify the theory.

In the end what matters is if your particular quirky old game works properly with a release.

I think one game I've seen problems with with drivers beyond 45.23 is NOLF if you want to try that. And of course there is valid concern with the driver overhead being higher in those newer drivers. And there's the question of whether there is really a benefit to using newer drivers beyond the magic version number being bigger.

If a newer forceware works good on a Geforce2, then you're running Windows XP. This thread clearly does not address Windows XP. It also doesn't matter how fast the CPU is either, it'll regress on anything.

The W9X forceware drivers are known to break on Geforce2s since 2004 or so. I've went through it first hand, it even breaks DX9 games. To nvidia, It's the way you're meant to be played - into buying their newer hardware. This was during their FX dark ages.

Pretty pathetic compared to ATI's final latest W9x drivers for Radeon7x00, and ATI's known for bad drivers... they've never topped it off with a string of broken drivers. (well until crimson for the non-GCN cards anyway)

45.23 seems to have a ton of fans here but after a lot of testing i found one specific issue with it: exiting winquake repeatably freezes the machine with a black screen, gf2 gts/440bx/98se. i've even bothered to downgrade to dx7.0a since that gets mentioned a lot, but no difference. neither 30.82 nor 61.72 had this issue.

another mention of issues with that version: https://www.tomshardware.com/reviews/aquamark3,683-9.html

also i think the doublescan option in nvidia drivers has only been added after 45.23.

Geforce always had Winquake issues (it's due to ddraw palettes and MGL) long before Forceware. You'll have to pass -dibonly to have it be stable.