[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mesa3d-dev
Subject:    Re: [Mesa3d-dev] Doom3 works on R200!
From:       Jon Trulson <jon () radscan ! com>
Date:       2004-10-25 18:56:39
Message-ID: Pine.LNX.4.61.0410251230490.20428 () m5 ! radscan ! com
[Download RAW message or body]

On Sun, 24 Oct 2004, Mike Mestnik wrote:

>
> --- Philipp Klaus Krause <pkk@spth.de> wrote:
>
>>
>>>>
>>>> Thank you for doing this work.  We really need to get the open-source
>>>> ATI driver on par with the propretary driver (both feature-wise and
>>>> performance-wise).
>>>
>>>
>>> But sadly we will NEVER match it.
>>>
>>> NO SmoothVision, HyperZ docu ever....
>>
>> The nonfree xig driver has been developed without HyperZ docs and
>> outperforms fglrx.
>>
> It's my opinion that xig uses trickery to get there FPS higher then it
> should be posible to under OpenGL compliant rendering.  It's also true in
> the font and 2d(like lack of xv support) accelerations as well.

 	Uhm... trickery?  Like how?  And what do you mean by 'should be 
posible to under OpenGL compliant rendering'?  Obviously we do it, so 
obviously it is possible.

 	If you mean extreme optimization of common code-paths, then you 
are correct.  If you are implying that we somehow 'cheat', well then you 
are quite incorrect.

 	We evaluate all drivers we produce.  If QA finds something amis 
(like really slow performance in some test that should be faster), then we 
spend time trying to figure out where the bottleneck is, and if we are 
successful in determining a cause, and it's a code-path worth optimizing, 
we deal with it.

 	I can assure you (though I don't think you will wish to believe 
me), there is no 'lets see if we are doing viewperf, and then cheat' code 
anywhere in our drivers.  I would think playing those kinds of games would 
be pretty obvious to detect... wouldn't they?  Didn't Nvidia get busted 
playing those games by some 3D testing outfit some time back?

 	Also, I am confused by your statements about lack of Xv, and 
'cheating' in 2D tests.

 	We do, of course, support Xv just fine (have since the first 
Summit release).  And how would you 'cheat' in a 2D test?  I mean, 
you can test the demo yourself and see that it does actually draw 
stuff...?  That's what Roland did I believe... I would think if Roland had 
found something suspicious, he would certainly have voiced them..?

[...]

PS: I work at XiG.  I do not subscribe to this list from there though, 
which is why I'm sending from home, where I am subscribed.  Feel free to 
flame me mercilessly :)

-- 
Jon Trulson    mailto:jon@radscan.com
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
"I am Nomad." -Nomad



-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
_______________________________________________
Mesa3d-dev mailing list
Mesa3d-dev@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
[prev in list] [next in list] [prev in thread] [next in thread]