Quadro - worth it? Any significant speed increases?
10119 5 2-
- eval
- Member
- 219 posts
- Joined: May 2008
- Offline
-
- Chrizto
- Member
- 66 posts
- Joined: Sept. 2008
- Offline
I use a Quadro FX card, but I don't know what you mean by speed increase…
The strengths of the professional GPU's lies in 1, the drivers, 2, the complete support for OpenGL and various supported attributes, like the ability to produce stereoscopic renders and stuff like that 3, the ability to use advanced programming API's to tailor your shader needs.
I would say, after using both consumer grade cards and both ATI FireGL and Nvidia Quadro's that you will definately notice a difference in how applications that relies on OpenGL support works. Everything just runs smoooth… ;-)
If it's worth the extra cash, well that's up to you to decide.
If you are going to purchase a Quadro card, (I would reccomend that vs ATI) you could go for a Quadro FX 1700 with 512MB RAM. The latest generation of this card has full support for the latest shader models and OpenGL, plus Cuda and all that.
You might say that the Quadro series is THE reference for OpenGL support in the GPU market today.
The FX 1700 is a mid-range card, with great price/performance.
Read all about it:
http://www.nvidia.com/page/qfx_mr.html [nvidia.com]
The strengths of the professional GPU's lies in 1, the drivers, 2, the complete support for OpenGL and various supported attributes, like the ability to produce stereoscopic renders and stuff like that 3, the ability to use advanced programming API's to tailor your shader needs.
I would say, after using both consumer grade cards and both ATI FireGL and Nvidia Quadro's that you will definately notice a difference in how applications that relies on OpenGL support works. Everything just runs smoooth… ;-)
If it's worth the extra cash, well that's up to you to decide.
If you are going to purchase a Quadro card, (I would reccomend that vs ATI) you could go for a Quadro FX 1700 with 512MB RAM. The latest generation of this card has full support for the latest shader models and OpenGL, plus Cuda and all that.
You might say that the Quadro series is THE reference for OpenGL support in the GPU market today.
The FX 1700 is a mid-range card, with great price/performance.
Read all about it:
http://www.nvidia.com/page/qfx_mr.html [nvidia.com]
“If your life is not NOW, you're already dead…”
From Chrizto's book of truths
From Chrizto's book of truths
-
- anon_user_40689665
- Member
- 648 posts
- Joined: July 2005
- Offline
-
- Remoth
- Member
- 2 posts
- Joined: April 2009
- Offline
I agree with everything cpb said, lol. The top end geoforce 200 series card is the same card as the top 5000 quadro series, just with more ram and more optimized drivers. For your own workstation at home, dont bother getting the Quadro, but if some client or boss offers to get it for you…. how can you say no? :twisted:
I've heard of some instances where people have tricked the PC into thinking that the geoforce card is a quadro card and getting the quadro drivers to work, thus giving them a cheap quadro. But you'd have to look that one up.
I've heard of some instances where people have tricked the PC into thinking that the geoforce card is a quadro card and getting the quadro drivers to work, thus giving them a cheap quadro. But you'd have to look that one up.
-
- probbins
- Member
- 1145 posts
- Joined: July 2005
- Offline
Much of this is going to depend on what you intend concentrating on with Houdini.
Are you going to do general type stuff that involves all areas of the app.?
Lots of really heavy modeling?
Lots of fluid simulations, particles?
For home use the only place you'll really have an advantage with a quad card is heavy modeling, meaning lots and lots of polygons. The quad card will handle the polygons in the viewport somewhat better.
Mind you, I'm using a nvidia 7150go on my laptop and can manipulated a 50+mb model without problems or lags.
What you do want is to stuff your computer with as much ram as it can stand, and that's the cheapest means to increase performance.
Then, for simulations you need cpu horsepower, the more the merrier.
For our type of work, particularly on a single machine, the graphics card is just a means of displaying what we are working on and rendering to the viewport in a decent colour range.
So go with the consumer card, it doesn't even have to be the latest and greatest. Use the saved money for more of the other things I mentioned.
Here's another recommendation. If you want to advance your workflow, spend $100 or 2 or 3 and buy extra Apprentice HD licenses (for the render tokens) and setup an HQueu distributed rendering and simulation pipeline. That's if you have an extra machine or two. They don't need much of anything regarding graphics cards, but again lots of ram is good.
Are you going to do general type stuff that involves all areas of the app.?
Lots of really heavy modeling?
Lots of fluid simulations, particles?
For home use the only place you'll really have an advantage with a quad card is heavy modeling, meaning lots and lots of polygons. The quad card will handle the polygons in the viewport somewhat better.
Mind you, I'm using a nvidia 7150go on my laptop and can manipulated a 50+mb model without problems or lags.
What you do want is to stuff your computer with as much ram as it can stand, and that's the cheapest means to increase performance.
Then, for simulations you need cpu horsepower, the more the merrier.
For our type of work, particularly on a single machine, the graphics card is just a means of displaying what we are working on and rendering to the viewport in a decent colour range.
So go with the consumer card, it doesn't even have to be the latest and greatest. Use the saved money for more of the other things I mentioned.
Here's another recommendation. If you want to advance your workflow, spend $100 or 2 or 3 and buy extra Apprentice HD licenses (for the render tokens) and setup an HQueu distributed rendering and simulation pipeline. That's if you have an extra machine or two. They don't need much of anything regarding graphics cards, but again lots of ram is good.
“gravity is not a force, it is a boundary layer”
“everything is coincident”
“Love; the state of suspended anticipation.”
“everything is coincident”
“Love; the state of suspended anticipation.”
-
- mrice
- Member
- 89 posts
- Joined: April 2008
- Offline
probbins
Here's another recommendation. If you want to advance your workflow, spend $100 or 2 or 3 and buy extra Apprentice HD licenses (for the render tokens) and setup an HQueu distributed rendering and simulation pipeline.
Is it possible to run hqueue with an HD license? It wont run here without a commercial license.
About the quadro, I doubt you'll see any benefit with Houdini as compared to newer geforce cards. Some apps (3dsmax) are much more stable with the quadro drivers so the cost is well worth it imo.
I dont think you can flash quadro bios on a geforce card anymore, since the 6800.
-
- Quick Links