PDA

View Full Version : Ati FireGL / Nvidia Quatro's - whats the big deal?


tfg
10-18-2004, 07:37 PM
This might seem like a stupid question, but since I'm not an animator I don't know much about professional 3d accelerators.

I'm getting together a small production for an animated film, and I'm interested helping my animators speed up their rendering times. I've read small bits about these "professional" accelerator cards such as the Ati FireGL. Can someone please explain how those cards speed up applications such as Maya/3dsMAX with some specifics: do they speed up the viewport window, or the final render process (or both)? What kind of elements do they typically accelerate inside those apps? Any personal opinions/stories would also be welcome.

Thanks!
Daniel :)

Vertizor
10-18-2004, 08:26 PM
Seaching the forums might help ;) This topic has been discussed many times over. But here's a quick recap of how it usually goes:

- These high end video cards only accelerate the viewport.

- No they do not speed up final rendering*

- * But certain 3D Labs cards coupled with drivers for specific 3D apps will offload final rendering to the GPU. Right now it's still not quite common practice.


So why do high end video cards matter? They move millions and millions of vertices... really fast... and accurately. The last part (accuracy) is very important to mechanical engineers and CAD people. If you're strictly animating, textures don't really matter at this point, you just want to see the geometry move. Typical gaming video cards put more emphasis on memory bandwidth to push more data and larger textures.

Think of all the polygons that go into making a humanoid face, then think of morphing them for facial animation, the feedback you get from pulling sliders and response time. Animating the proxy cage on a sub divided surface is a shortcut instead of trying to animate millions of polys, but it has potential for undesired results.

My personal experience, back in the day when the original GeForce one came out, and the Quadros of that time were really just GeForce 3 cores with better drivers, and this was before the GF3 core became mainstream. I had bought a Elsa Gloria. It had 64 Mb of V RAM and that was a big deal back then because video cards typically had 32 MB. That Quadro card could move polys like nothing I've seen before. And it cost me $800. Today, a $200 video could out beat that Quadro, but the workstation cards also move up in evolution.

Tarrbot
10-20-2004, 02:34 AM
Vertizor, I know the feeling. I had an Elsa GLoria as well. It finally bit the dust last year.

In the day, however, it was a great card for a great price. I've owned several Elsa video cards and have been impressed by them all.

Vertizor
10-20-2004, 02:52 AM
Yeah, you really won't appreciate it until you own one... and crank up the mesh smooth... and spin her around a couple times.

Come to think of it, I wonder if the drivers do any vertex caching irregardless of what the software code is doing. Maybe that's what makes those drivers so special.

CGTalk Moderation
01-19-2006, 02:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.