If Nvidia gets the Fermi issues ironed out; It will be interesting to see how the Fermi effects the QuadroFX lineup. Will we see a Fermi based QuadroFx?
Absolutely, sir. In fact, Fermi's design is very much oriented towards accelerating real-time rendering and GPU compute, two things that are prevalent in programs like Maya, Lightwave, etc.
Now we just need the software vendors to start building their programs to really take advantage of these higher end video cards. To date Autodesk still isn't certifying hardware for Autodesk Revit products, and Adobe is only certifying parts of their product line up. There is a lot of change happening thanks to these newer cards and GP2 abilities.
There are a couple pieces of software that are starting to take advantage of GPU compute. Adobe Premiere CS4 offers GPU-acceleration on both NVIDIA and AMD. NVIDIA's solution only accelerates CS4 on Quadro cards, while ATI's implementation does it on all Radeon HD 2000+ GPUs.
I do know that AMD also works closely with Autodesk, with the latter certifying AMD's FirePro adapters which allows the customer direct contact with an Autodesk software engineer for support. That's not GPU compute, however. I'll see if I can get Bobby, our DCC guy, in here to talk a little more about what 3D DCC firms are doing.
Fermi is both the most exciting thing currently in development, and the most terrifying. It has potential to be an incredible work of hardware, but it could also fail harder than anything NVIDIA's likely seen. They're putting a lot of faith in Fermi, and if it doesn't deliver, things will get really ugly.
Regarding what Walt said - I can't state any specifics, but you can be assured that AMD certianly does have the upper hand with gaining certifications. When I spoke with them at SIGGRAPH three months ago, they mentioned that they were making great efforts to work with software developers to optimize and certify with FirePro. We'll most likely see progression in any area with FirePro before we do with Quadro.
But you are right, the face of potential is changing. If everyone rolls with the punches, we'll see incredible performance yields in 3D apps and other professional graphics applications.
Don't get your hopes up for adobe though, at least not immediately. They tend to buck change frequently, and I don't see this situation being any different.
With the internet and especially the emergence of social media, corporate transparency is becoming a consumer expectation.
I understand that NVIDIA has investors that count on continued progress reports, and they have potential customers that will hold off on purchasing rival product if they think the latest and greatest is just around the bend, but I still have trouble seeing how this is a healthy approach long term.
For the sake of their own credibility I think they would be far better served by coming out and saying hey, its a really ambitious project, one that will be worth the wait. Instead they get the CEO to grin next to a mock up leading many to believe Fermi was in the bag for holiday 2009. I'm not sure anyone will be able to get their hands on one before March.
Comments
I do know that AMD also works closely with Autodesk, with the latter certifying AMD's FirePro adapters which allows the customer direct contact with an Autodesk software engineer for support. That's not GPU compute, however. I'll see if I can get Bobby, our DCC guy, in here to talk a little more about what 3D DCC firms are doing.
Regarding what Walt said - I can't state any specifics, but you can be assured that AMD certianly does have the upper hand with gaining certifications. When I spoke with them at SIGGRAPH three months ago, they mentioned that they were making great efforts to work with software developers to optimize and certify with FirePro. We'll most likely see progression in any area with FirePro before we do with Quadro.
But you are right, the face of potential is changing. If everyone rolls with the punches, we'll see incredible performance yields in 3D apps and other professional graphics applications.
Don't get your hopes up for adobe though, at least not immediately. They tend to buck change frequently, and I don't see this situation being any different.
I understand that NVIDIA has investors that count on continued progress reports, and they have potential customers that will hold off on purchasing rival product if they think the latest and greatest is just around the bend, but I still have trouble seeing how this is a healthy approach long term.
For the sake of their own credibility I think they would be far better served by coming out and saying hey, its a really ambitious project, one that will be worth the wait. Instead they get the CEO to grin next to a mock up leading many to believe Fermi was in the bag for holiday 2009. I'm not sure anyone will be able to get their hands on one before March.
GPU processing truly is the future. It's becomeing the focus of everyone.