No More Shader Replacement

The secret is all in compilation and scheduling. Now that NVIDIA has had more time to work with scheduling and profiling code on an already efficient and powerful architecture, they have an opportunity. This generation, rather than build a compiler to fit hardware, they were able to take what they've learned and build their hardware to better fit a mature compiler already targeted to the architecture. All this leads up to the fact that the 7800 GTX with current drivers does absolutely no shader replacement. This is quite a big deal in light of the fact that, just over a year ago, thousands of shaders were stored in the driver ready for replacement on demand in NV3x and even NV4x. It's quite an asset to have come this far with hardware and software in the relatively short amount of time NVIDIA has spent working with real-time compilation of shader programs.

All these factors come together to mean that the hardware is busy more of the time. And getting more things done faster is what it's all about.

So, NVIDIA is offering a nominal increase in clock speed to 430MHz, just a little more memory bandwidth (256bit memory buss running at a 1.2GHz data rate), 1.33x vertex pipelines, 1.5x pixel pipelines, and various increases in efficiency. These all work together to give us as much as double the performance in extreme cases. If the performance increase can actually be realized, we are looking at a pretty decent speed increase over the 6800 Ultra. Obviously, in the real world we won't be seeing a threefold performance increase in anything but a bad benchmark. In cases where games are CPU limited, we will likely see a much lower increase in performance, but performance double that of the 6800 Ultra is entirely possible in very shader limited games.

In fact, EPIC reports that under certain Unreal Engine 3 tests they currently see two to 2.4x improvements in framerate over the 6800 Ultra. Of course, UE3 is not finished yet and there won't be games out based on the engine for a while. We don't usually like reporting performance numbers from software that hasn't been released, but even if these numbers are higher than we will see in a shipping product, it seems that NVIDIA has at least gotten it right for one developer's technology. We are very interested in seeing how next generation games will perform on this hardware. If we can trust these numbers at all, it looks like the performance advantage will only get better for the GeForce 7800 GTX until Windows Graphics Foundation 2.0 comes along and inspires new techniques beyond SM3.0 capabilities.

Right now, each triangle that gets fed through the vertex pipeline, there are many pixels inside the object that needs her help.

Bringing It All Together

Why didn't NVIDIA build a part with unified shaders?

Every generation, NVIDIA evaluates alternative architectures, but at this time they don't feel that a unified architecture is a good match to the current PC landscape. We will eventually see a unified shader architecture from NVIDIA, but it will not likely be until DirectX itself is focused around a unified shader architecture. At this point, vertex hardware doesn't need to be as complex or intricate as the pixel pipeline. As APIs develop more and more complex functionality it will be advantageous for hardware developers to move towards a more generic and programmable shader unit that can easily adapt to any floating point processing need.

As pixel processing is currently more important than vertex processing, NVIDIA is separating the two in order to focus attention where it is due. Making hardware more generic usually makes it necessarily slower, but explicitly targeting a specific aspect of something can often improve performance a great deal.

When WGF 2.0 comes along and geometry shaders are able to dynamically generate vertex data inside the GPU we will likely see an increased burden on vertex processing as well. Being able to programmatically generate vertex data will help to remove the burden on the system to supply all the model data to the GPU.

Inside The Pipes Transparency AA, Purevideo, and HDTV
Comments Locked

127 Comments

View All Comments

  • multiblitz - Sunday, June 26, 2005 - link

    It would be great of you could do a comparison between the 6800 and the 7800 in video /DVD-playback-quality similar to the comparison betwenn the X800 and the 6800 you did last year.
  • at80eighty - Saturday, June 25, 2005 - link

    OMG! I've never seen so many bitching whiners come outta the woodworks like this!!

    You A-holes oughta remember that this site has been kept free

    F
    R
    E
    E

    The editors owe YOU nothing. At all.

    AT team - accidents happen. Keep up the great work!

    /#121 : well said. Amazing how these turds dont realise that the knife cuts both ways...
  • mrdeez - Friday, June 24, 2005 - link

    #124
    You can stfu too...j/k..point taken .

    I guess the real issue for me is that this card is a beast but ill never have it in my sli rig......i want all settings maxed at playable resolutions thats just me.........and i will not go back to crt...lol crt thats was lame dude
  • Momental - Friday, June 24, 2005 - link

    #122 The problem with your solution regarding "all of us just getting two 6880U's" works perfectly for those with an SLI-capable board, yes? Some of us, like myself, anticpated the next generation of GPU's like the 7800 series and opted to simply upgrade to one of those when the dust settled and prices slid back a bit.

    Additionally, telling someone to "STFU" isn't necessary. We can't hold a conversation if we're all silent. Knowhuddamean, jellybean? Hand gestures don't work well over the internet, but here's one for you..........
  • SDA - Friday, June 24, 2005 - link

    LCD gamers shouldn't be bothering with new graphics cards, they should get new monitors.

    kidding, I have nothing against LCDs. The real advantage of showing the card run at 2048x1536 is that it lets you see how well the card scales to more stressful scenarios. A card that suddenly gets swamped at higher resolutions will probably not fare well in future games that need more memory bandwidth.

    On a side note, you can get a CRT that will run 2048x1536 @ a reasonable refresh for about $200 shipped (any Sony G520 variant, such as the Dell P1130). The only things that would actually be small in games are the 2D objects that have set pixel sizes, everything else looks beautiful.
  • mrdeez - Friday, June 24, 2005 - link

    #121
    lol ty for your insight....anyway like i said this card is not for lcd gamers as most have a 12x10 or 16x12.....so what purpose does this card have??answer me this batman and you have the group that should buy this card -otherwise, the rest of us should just get 2 6800u....this card is geared more for workstation graphics not gaming.....unless you game on a hi def crt and even then max res would be 1920 by 1080i..or something like that.....
  • SDA - Friday, June 24, 2005 - link

    #116, if people in the comments thread are allowed to give their opinions, why shouldn't #114 give his too? Surely even an illiterate like you should realize that arguing that everyone is entitled to his or her own argument means that the person you're arguing with is too.

    #119, some people have different requirements than others. Some just want no visible blur, others want the best contrast ratio and color reproduction they can get.
  • bob661 - Thursday, June 23, 2005 - link

    #188
    Oh yeah. The monitor goes up to 16x12.
  • bob661 - Thursday, June 23, 2005 - link

    #118
    I play BF2 on a Viewsonic VP201b (20.1") at work and it's very good. No streaking or ghosting. Video card is a 6800GT. I play at 1280x960.
  • Icehawk - Thursday, June 23, 2005 - link

    Well, I for one think 1280x1024 is pretty valid as that is what a 19" LCD can do. I'd just want to see a maxed out 12x10 chart to see how it does - I know a 6800 can't do it for every game with full AA and AF. Otherwise I agree - a 12x10 with no options isn't going to show much with current games.

    See, I'm considering swapping my two 21" CRTs for two 19" LCDs - and they won't do more than 12x10. I'd love to do two 20-21" LCDs but the cost is too high and fast panels aren't to be found. 19" is the sweet spot right now IMO - perhaps I'm wrong?

    Thanks AT for a nice article - accidents happen.

Log in

Don't have an account? Sign up now