Hi Dirk.

>> BTW, I'm going all the way this time for a graphics card.  The decisioin 
>> has come down to a Quadro 5600 or FireGL 8650.  Based on 
>> 3dProfessor.com's review, I'm leaning toward the FireGL.  It has 2 GB 
>> versus 1.5 GB for the Quadro, and the benchmarks are superior in several 
>> ways.  Does everyone agree?  I'm not sure, because another review of the 
>> same two cards favored the Quadro in all tests, save two (Maya and 
>> Perview).  Also, this second review complained about fan noise in the 
>> FireGL.  Does anyone have an opinion on high-end graphics cards?
>
> I haven't used ATI cards in a long time, so take the following with a 
> grain of salt (I just won a little one at GDC, so I might give them 
> another try when I find some time ;). The main problem with ATI cards used 
> to be the drivers. Their OpenGL support was noticeably worse than 
> nVidia's, especially on Linux. The performance was pretty good, last time 
> I tested, but stability and bugs were issues.
>
> I don't know your models, but 1.5G is a lot of memory. If you fill that up 
> (especially if you don't have large textures), you're pushing the 
> capabilities of the card for interaction already, and I don't think the 
> additional 500 MB are going to make a big difference.

I wanted to have the extra memory to have the option of doing CUDA 
programming directly on the card, in which case the extra memory would be 
used for model data, too, which would spill over into system memory.  I'm 
just now looking into the CUDA, but this is exclusively an nVidia 
technology.  I don't know how useful it is yet for my particular problem, 
and I don't know what ATI's answer to CUDA is on their FireGL cards.

>
>> I could save a lot a money and go with an 8800 Ultra or HD 3870 X2, but I 
>> think the texturing and AA accuracy would suffer in favor of the speed, 
>> and I'm not really looking for high-frame rates, if I cannot also produce 
>> frames of cinemagraphic quality.  I'm making high-quality simulations and 
>> movies. I don't understand why the driver can't be configured or changed 
>> so that the workstation versions of these cards can go faster but at 
>> lower graphical detail/rendering accuracy.  Reversely, I'm not convinced 
>> that an 8800 Ultra can be made to perform like a Quadro 5600 with some 
>> driver tweaks, but I'm willing to learn how to do this if it is possible. 
>> Does anyone here do this?
>
> IMHO, the professional cards don't really make better pictures than the 
> gamer cards. The main difference on nVidia is performance for lines, and 
> the ability to run stereo/genlock/sync.

The lines are better because the AA is better?  I thought texturing worked 
better on the workstation cards, too, but I don't know the details.  I'm 
trying to find someone who does enough fine cinema work and gaming to know 
whether the workstation cards are really worth it.

>
> If you want cinematic quality at the cost of speed you can always combine 
> multiple images for more samples.

I'm not sure how the multiple-images technique works.  I was thinking in 
terms of how AA and texture-superpositioning are done, and perhaps also 
advanced lighting configurations.
>
>> Do most of you on Windows (XP Pro SP2/3) prefer to use Visual Studio .NET 
>> for your code development?  I'm going to wait for the new machine before 
>> I put VS back on.  So, I'll wait until then (two weeks or so) until I 
>> begin looking at source.
>
> Can't comment on that one, I don't use Windows on a daily (or monthly ;) 
> basis.

You are fortunate.  :)


Shaping 


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to