The following is what I wrote fellow webmasters of which I'm waiting for confirmation :
If I get confirmation of my findings, I'm sure some website(s) will report this.Guys,
Don't worry this is not a email asking for a plug :smile:.
I hope you will all have time for an experiment. This has all got to do with NVIDIA's GeForce4 MX440 and a recent finding of mine. NOTE that all of this is based on NV drivers ver 27.30 on Win98SE but I have a good feeling it should work with all 27.xx drivers and possibly WinXP.
As you all know, NVIDIA uses a Unified Driver Architecture. Everything I say below has to do with **switching an existing GF3 (any GF3 iteration) to a GF4 MX440**.
Method A, which is the "correct method", of switching from a GF3 to a GF4 MX :
1) Uninstall existing video card drivers - reboot
2) Upon reboot, standard Windows PCI VGA drivers are asked to be loaded and used - do as told and again a reboot is required
3) Upon reboot, standard PCI VGA is loaded and used; you shut down and switch cards and boot again
4) Upon boot, you are asked for the drivers of the new card; so you point to the right direction (you know what I mean); reboot again required
5) On this new boot, everything should work as intended with the GF4 MX440 recognized as the vid card in the system
Method B : Now, try this instead of the above when switching from a GF3 to a GF4 MX440 :
1) Instead of uninstalling existing NV drivers are per (1) above, simply shutdown and switch your GF3 to the GF4 MX440; reboot
2) Upon boot, Windows should recognize the GF4 MX40 as such, loads appropriate identification entries *and* uses the same/existing 27.xx drivers and then asks for a reboot; reboot
3) Upon reboot, GF4 MX440 is recognized as existing card
Now, try Method A for switching from a GF3 to a GF4 MX440 and run any benchmark/game *without AA* on the GF4 MX440. Record results.
Next try Method B for switching from a GF3 to a GF4 MX440 and run the same benchmark/game *without AA* on the GF4 MX440. Record results.
Tell me if Method B does not result in *significantly* higher results than Method A, because that's the way it is for me - in F1 2001 (D3D), at 1024x768x32 NoAA, Method A gets me 25fps while Method B gets me 49fps!!! . Note that my observations are based on a single game, EA's D3D F1 2001. Using Method B also results in all of the NV AA not working at all.
I'm not sure if this of any value but perhaps for those that do not care about AA at all, it may if the GF4 MX440 is at the budget they're in.
Of course, all of this is for nothing if you don't try my above and get the same basic result.
The point is, of course, having a GF4 MX440 could possibly result in performances higher than expected through the "direct switch from a GF3 to a GF4 MX440" instead of the "correct" method of upgrading from a GF3 to the GF4 MX440.
If you have the time to try this out and can confirm my findings, feel free to make your own article for the public regarding this.And of course, my system may be completely screwed up... but I honestly
don't think so :smile:.
Cheers and hoping you will have the time to try this out, prove that it is consistent with my finding and let the world know the GeForce4 MX440 may be purposely "crippled" by NVIDIA... but maybe not through a driver
and OS loophole.