Inside The Pipes

The pixel pipe is made up of two vector units and a texture unit that all operate together to facilitate effective shader program execution. There are a couple mini-ALUs in each shader pipeline that allow operations such as a free fp16 normalize and other specialized features that relate to and assist the two main ALUs.



Even though this block diagram looks slightly different from ones shown during the 6800 launch, NVIDIA has informed us that these mini-ALUs were also present in NV4x hardware. There was much talk when the 6800 launched about the distinct functionality each of the main shader ALUs had. In NV4x, only one ALU had the ability to perform a single clock MADD (multiply-add). Similarly, only one ALU assisted in texture address operations for the texture unit. Simply having these two distinct ALUs (regardless of their functionality difference) is what was able to push the NV4x so much faster than the NV3x architecture.

In their ongoing research into commonly used shaders (and likely much of their work with shader replacement), NVIDIA discovered that a very high percentage of shader instructions were MADDs. Multiply-add is extremely common in 3D mathematics as linear algebra, matrix manipulation, and vector calculus are a huge part of graphics. G70 implements MADD on both main Shader ALUs. Taking into account the 50% increase in shader pipelines and each pipe's ability to compute twice as many MADD operations per clock, the G70 has the theoretical ability to triple MADD performance over the NV4x architecture (on a clock for clock basis).

Of course, we pressed the development team to tell us if both Shader ALUs featured identical functionality. The answer is that they do not. Other than knowing that only one ALU is responsible for assisting the texture hardware, we were unable to extract a detailed answer about how similar the ALUs are. Suffice it to say that they still don't share all features, but that NVIDIA certainly feels that the current setup will allow G70 to extract twice the shader performance for a single fragment over NV4x (depending on the shader of course). We have also learned that the penalty for branching in the pixel shaders is much less than in previous hardware. This may or may not mean that the pipelines are less dependent on following the exact same instruction path, but we really don't have the ability to determine what is going on at that level.

No More Memory Bandwidth No More Shader Replacement
Comments Locked

127 Comments

View All Comments

  • swatX - Wednesday, June 22, 2005 - link

    THE SLI is meant to played on high res.. if you got money to brn on SLI then i am damn sure you got money to burn on a 19" monitor ;)
  • CtK - Wednesday, June 22, 2005 - link

    can Dual Display be used in SLi mode??
  • Johnmcl7 - Wednesday, June 22, 2005 - link

    In general 6600GT SLI performance seems a bit random, in some cases it's really good as with BF2 but in others not as good as a 6800GT.

    John
  • bob661 - Wednesday, June 22, 2005 - link

    Anyone notice how a SLI'd 6600GT is just as quick as a 6800 Ultra in BF2?
  • R3MF - Wednesday, June 22, 2005 - link

    give me some details on the 7800 and 7800GT

    what, when, and how much?
  • bob661 - Wednesday, June 22, 2005 - link

    #59
    I am more eager to see how the new midrange cards will perform than these parts but if I had a spare $600 I would jump all over this.
  • bob661 - Wednesday, June 22, 2005 - link

    #56
    LMAO!!!!
  • bob661 - Wednesday, June 22, 2005 - link

    #44
    And I thought paying $350 for a video cards was too much then or even before than there was the $200 high end and before that the $100 high end. I balked at all of those prices but I understood why they were prices as such and didn't bitch everytime the costs went up. The bar keeps being raised and the prices go with it. Inflation, more features and the fact that most of us here can afford $350 video cards pushes the cost of new PREMIUM cards higher by the year. It's only going to go up unless either people quit buying the high end cards or the manufatucrers find a magical process to reduce costs dramatically.
  • Johnmcl7 - Wednesday, June 22, 2005 - link

    You're quite right, there's always a premium for the best, I don't see any difference here, no-one is being forced to buy this graphics card. As usual, I'll wait until something offers me a better price/performance ratio over my current X850XT/6800 Ultra duo.

    John
  • Avalon - Wednesday, June 22, 2005 - link

    Seems to be a problem with the last Knights of the Old Republic 2 graph. Both 7800GTX setups are "performing" less than all the other cards benched. Despite all the mistakes, it still seems like I was right in that this card is made for those who play at high resolutions. Anyone with an R420 of NV40 based card that plays at 16x12 or less should probably not bother upgrading, unless they feel the need to.

Log in

Don't have an account? Sign up now