NVIDIA and ATI Technologies, the two largest GPU vendors are currently dueling over the next generation GPU chipsets.
NVIDIA is currently proceeding with the launch of the GeForce 7800 GTX (G70), while ATI prepares its next generation GPU codenamed "R520". NVIDIA's product roadmap includes the launch of the “G71 (old NV49)" as a mid-range product, available in the end of 3Q 2005, and the main stream "G72 (old NV46)", scheduled for Q4 2005. On the other hand, after the launch of the R520 (Autumn 2005), Canadian ATI prepares the "RV530" for mid range and the "RV515" main stream and more affordable solution.
Meanwhile, NVIDIA is currently the partner of Sony in the development of the upcoming PS3 gaming machine. Available in March 2006, the PS3 will use an NVIDIA-designed graphics processor called RSX. Both RSX and G70 GPUs share the same heritage, and same technology. The RSX, however is expected to be faster.
On the other hand, ATI will power up the graphics of the next-generation game console of Microsoft, the XBOX 360, with the R500 graphics chipset (90 nm process). Latest information however indicates that ATI will not be the actual manufacturer of the R500 chips, but it will just design them. IBM will be possibly the actual manufacturer of the Xbox 360 GPU. Microsoft plans to ship Xbox 360 consoles world wide starting from October 2005.
NVIDIA’s RSX for PS3 is claimed that it will carry 300 million transistors, more than what the R500 is expected to have but less than what the R520 will feature.
A basic difference between the two graphics giants is their different approach in their developing architecture. NVIDIA chose to work on the current development route, based on the DirectX 9 API and the 110nm process. Their work was mostly a "refine" of the hardware, resulting however, to significantly improved performance. On the other hand, ATI has chosen to reform the structure of its GPU chipset from root. The ATI R520 GPU will be manufactured by using the 90nm process, a decision that has possibly forced ATI to delay the launch of its R520. Should ATI finally succeed in its 90nm developments, meaning that the new architecture will work efficiently in the production table, the possibility for the company to further strengthen and extend will be very high.
ATI and the Unified-Shader architecture
Graphics are generally made up by vertices and pixels. Both should be shaded by the graphics chip, with two different mechanisms. These are the vertex pipelines and the pixel pipelines. The operation of these pipelines is controlled by both hardware and software. Regarding the software layer, the possibility to use a common or unified layer to control both vertex and pixel pipelines is feasible. A unified shader architecture is currently a debate in the graphics industry, whether or not a Unified Shader pipeline is better.
NVIDIA chose to independently control the Vertex and Pixel shaders, with its G70 GPU, a reformed and improved version of the NV40. By expanding the number of shaders as well as the efficiency of the pipelines, NVIDIA has managed to increase the overall shader performance by 50%.
On the other hand, ATI has already revealed that it will adopt the Unified-Shader architecture in the development of the Xbox 360 GPU. According to the Unified-Shader concept, no distinct dedicated vertex and pixel shader engines are used. Instead these are replaced by an array of Unified-Shader engines able to execute both kinds of instructions.
ATI believes that such an approach offers the best performance and also better allocation of the GPU resources, compared to NVIDIA’s dual shaders. NVIDIA’s answer to these claims is that the Unified-Shader technology could be also a
future option for the company, as long as it will be assured that its operation will be smooth, easily controlled and predictable.
The Unified-Shader architecture is currently backed by Microsoft, as it is already included in the next-generation Xbox 360 game console, and it might be a part of the new graphics API for Windows Longhorn – Windows Graphics Foundation 2.0.
Windows Graphics Foundation will be the graphics subsystem of the Longhorn OS, expected in the market in 2006. The WGF2.0 includes the DirectX 10 or and also Shader model 4.0 support. ATI seems to anticipate WGF2.0 with the R5xx series of GPUs. But NVIDIA comments that the WGF2.0 is an API (software) and does not necessarily point out to the Unified-Shader concept.
NVIDIA Corp.’s chief architect David Kirk said in an interview with Bit-tech.net web-site:
"Well, let's get something straight. Microsoft makes APIs (Application Programming Interfaces- Ed) not hardware. WGF is a specification for an API specification - it's software, not hardware."
"For them (Microsoft), implementing Unified Shaders means a unified programming model. Since they don't build hardware, they're not saying anything about hardware."
"Debating unified against separate shader architecture is not really the important question. The strategy is simply to make the vertex and pixel pipelines go fast. The tactic is how you build an architecture to execute that strategy. We're just trying to work out what is the most efficient way."
"It's far harder to design a unified processor - it has to do, by design, twice as much. Another word for 'unified' is 'shared', and another word for 'shared' is 'competing'. It's a challenge to create a chip that does load balancing and performance prediction. It's extremely important, especially in a console architecture, for the performance to be predicable. With all that balancing, it's difficult to make the performance predictable. I've even heard that some developers dislike the unified pipe, and will be handling vertex pipeline calculations on the Xbox 360's triple-core CPU."
The next generation of NVIDIA’s GPU will be the "G80 (NV50)". It is not clear whether the G80 will feature a Unified-Shader architecture, or an independent formation Shader.