Nvidia’s Transition to Unified Shader Architecture Will Be Gradual – Chief Scientist

Nvidia Set to Take ATI-Like Approach to Unified Shader Architecture

by Anton Shilov
04/18/2006 | 05:08 PM

Nvidia Corp.’s chief scientist David Kirk has once again reaffirmed his stance on the unified shader processors by claiming that shortly from now there may be no point to implement them and said that the company would be “gradually” going to the unified architecture, the approach taken by arch-rival ATI Technologies.


“The vertex shader, pixel shader, ROP (rasterizing operation), tesselator, other hardware units [of the contemporary graphics processors] will probably evolve into something that can execute everything. But time is required for, [such evolution] is not possible at one time. Change probably will happen progressively,” said David Kirk in an interview with Japanese PC Watch web-site (please note that the quote was translated from Japanese – X-bit labs).

Previously Nvidia Corp. expressed opinion that it would be necessary to release a micro-architecture with unified shader processors “when it makes sense”, alluding that it may not be the company’s first design that supports DirectX 10 and Shader Model 4.0, which takes advantage of unified shader processors, which can execute different types of shaders, including pixel and vertex. Nevertheless, Nvidia’s David Kirk has never said which of the company’s products will sport and which will not unified shader processors.

The modern rendering pipeline from Nvidia consists of vertex shader processors, texture units, pixel shader processors ROPs. Even though the latter act nearly independently these days, the graphics pipeline of Nvidia’s GeForce 7-series hardware is still streamlined, something which DirectX 10 ideology wants to get rid off. The new philosophy, which is generally shared by ATI, Microsoft and Nvidia, claims that engines that perform shaders should be unified. Moreover, ATI has split up the pixel pipeline totally in the most recent Radeon X1000-series: pixel shader processors, texture units and ROPs act totally independently and their workload is managed by a special arbiter processor, which is ideologically close to the DirectX 10 graphics chips, something which ATI’s Richard Huddy called “gradual” architectural transition to DirectX 10 back in 2004 and something which will be adopted by Nvidia Corp. too.

Earlier published rumours suggest that Nvidia’s code-named G80 graphics processing unit (GPU) will incorporate 48 pixel shader processors and an unknown number of vertex shader processors, some unofficial sources said. The chip is still expected to support feature-set of DirectX 10 along with Shader Model 4.0, even though it will not take advantage of the unified shader processors that can compute both pixel and vertex shaders. The architecture of the G80 (or the GeForce 8800, if the company decides to follow its current nomenclature) should be totally different from the GeForce 7 (G70, G71) or the Radeon X1000 (R520, R580), however, some ideology may be similar to that of ATI’s code-named R580.

Nvidia Corp. does not comment on unreleased products..