feedback to NVidia [was: Nvidia Drivers]

[Greg Trounson]

Frank Tanner III wrote:
> --- Sean Estabrooks <seanlkml at sympatico.ca> wrote:
> 
>>On Wed, 26 May 2004 09:54:15 -0700 (PDT)
>>Frank Tanner III <pctech at mybellybutton.com> wrote:
>>
>>
>>>If they don't publish the specifications how are
>>
>>their
>>
>>>all of these third party shareware tweak programs
>>
>>for
>>
>>>the nVidia cards?  Easy answer.  Because there ARE
>>>published specs.  You're just too busy ranting to
>>>bother to look for them.  You'd rather be part of
>>
>>the
>>
>>>problem rather than part of the solution.
>>
>>Care to back this up with some evidence that isn't
>>quite so speculative?   Do you have a link for
>>specifications
>>of these cards?   AFAIK there aren't any available
>>to the
>>public, otherwise there'd almost certainly be open
>>source drivers.
>>
>>
>>Regards,
>>Sean
> 
> 
> As I stated in my e-mail to you directly, the
> existance of the tweak utilities themselves are the proof.

No, the existance of tweak utilities is merely proof that the registers 
that allow tweaking are documented.  For example GPU clock speed or 
memory bus speed.

And the fact that we have an open source 2D driver is pretty much 
evidence that the 'put a pixel on the screen' function is documented.

How these became documented I am not sure of - is it notes taken as a 
result of reverse-engineering, or did nVidia publish them at some point?

What is *NOT* documented is pretty much everything else, such as the 
method by which the driver should communicate with the GPU on the card 
to produce accelerated 3D graphics, or the registers that allow for 
multiple monitors.  In other words, all the reasons you bought that card 
in the first place instead of just plugging in your 1997 S3 Trio64 and 
getting the exact same performance.

Greg