ABOUT BENCH FORUMS PODCAST LOGIN REGISTER
PC Components
Smartphones & tablets
Systems
ENTERPRISE & IT
GUIDES

ATI Radeon HD 2900 XT: Calling a Spade a Spade

 by Derek Wilson on May 14, 2007 12:04 PM EST
86 Comments | Add A Comment
86 Comments

Texturing, Caches, and Memory

Texturing

R600 features less texture hardware than we would expect to see, though AMD stands by the argument that compute power will come out on top when it matters. At the same time, we can't compute anything if we don't have any data to work with. So let's take a look at what AMD has done with their texture units.

There are four texture units in R600, one for each SIMD unit. These units don't share resources with the hardware in the SIMD units and are independently scheduled by AMD's dispatch processor. The dispatch processor is able to determine what data will be needed for threads about to execute and can handle setting up the texture units without waiting for the SIMD unit to request data and come up empty.

Texture units on the R600 are able to make both filtered and unfiltered texture requests no matter what shader is running. Unfiltered textures are useful with non-image-based texture data like vertex textures, normal maps, and generic blocks of data. Filtered requests will generally be for image data to be used in determining the color of a pixel. R600 can address one unfiltered texture per clock per texture unit and one filtered textures per clock per texture unit. Filtered units can be used to request unfiltered textures if necessary, providing an extra four unfiltered textures in place of one filtered texture.

The unfiltered texture requests will come back through four fp32 texture samplers (one per component), while the filtered requests will return 16 data points which will be run through the texture filtering hardware resulting in four filtered texture samples. The hardware can at best produce 32 single component fp16 unfiltered results per texture unit per clock. More practically, each texture unit can produce four bilinear filtered four component fp16 samples per clock alongside four unfiltered results. For textures with fp32 components, two clocks would be required to complete a bilinear filter process, as only half the data is loaded at a time to conserve bandwidth.

This is definitely a step up for R600, as R5xx hardware doesn't have texture filtering hardware for floating point textures. All told, with each of its four texture units working, R600 can consume up to 32 unfiltered textures or 16 unfiltered textures plus 16 filtered textures (as long as they're fp16 or fewer bits and we're only using bilinear filtering).

G80 is built with four texture address units and eight texture filters per block of 16 SPs. In total, this means NVIDIA's hardware can produce 32 filtered texture samples per clock (again these are fp16 and bilinear filtered). Of course, NVIDIA is operating on twice as many threads per clock, so it is conceivable that they would benefit more from having the extra filtered data.

We will have to wait and see if AMD's approach of providing unfiltered and filtered texture access in parallel pays off. For the general case on pixel shaders, we would want to see more filtered textures per clock, but with vertex and geometry shaders coming into the mix this could be a good way to save hardware space while offering more texturing power. On a final texturing note, AMD implemented "percentage closer" filter hardware for depth stencil textures. This will allow developers to implement fast soft shadows. The details of the implementation weren't indicated though.

Next Up: NVIDIA's G80 Finally: A Design House Talks Cache Size
PRINT THIS ARTICLE
Comments Locked

86 Comments

View All Comments

  • Roy2001 - Tuesday, May 15, 2007 - link

    The reason is, you have to pay extra $ for a power supply. No, most probably your old PSU won't have enough milk for this baby. I will stick with nVidia in future. My 2 cents.

  • Chaser - Tuesday, May 15, 2007 - link

    quote:

    While AMD will tell us that R600 is not late and hasn't been delayed, this is simply because they never actually set a public date from which to be delayed. We all know that AMD would rather have seen their hardware hit the streets at or around the time Vista launched, or better yet, alongside G80.

    First, they refuse to call a spade a spade: this part was absolutely delayed, and it works better to admit this rather than making excuses.



    Such a revealing tech article. Thanks for other sources Tom.

  • archcommus - Tuesday, May 15, 2007 - link

    $300 is the exact price point I shoot for when buying a video card, so that pretty much eliminates AMD right off the bat for me right now. I want to spend more than $200 but $400 is too much. I'm sure they'll fill this void eventually, and how that card will stack up against an 8800 GTS 320 MB is what I'm interested in.

  • H4n53n - Tuesday, May 15, 2007 - link

    Interesting enough in some other websites it wins from 8800 gtx in most games,especially the newer ones and comparing the price i would say it's a good deal?I think it's just driver problems,ati has been known for not having a very good driver compared to nvidia but when they fixed it then it'll win

  • dragonsqrrl - Thursday, August 25, 2011 - link

    lol...fail. In retrospect it's really easy to pick out the EPIC ATI fanboys now.

  • Affectionate-Bed-980 - Tuesday, May 15, 2007 - link

    I skimmed this article because I have a final. ATI can't hold a candle to NV at the moment it seems. Now while the 2900XT might have good value, I am correct in saying that ATI has lost the performance crown by a buttload (not even like X1800 vs 7800) but like they're totally slaughtered right?

    Now I won't go and comment about how the 2900 stacks up against competition in the same price range, but it seems that GTSes can be acquired for cheap.

    Did ATI flop big here?

  • vailr - Monday, May 14, 2007 - link

    I'd rather use a mid-range older card that "only" uses ~100 Watts (or less) than pay ~$400 for a card that requires 300 Watts to run. Doesn't AMD care about "Global Warming"?
    Al Gore would be amazed, alarmed, and astounded !!

  • Deusfaux - Monday, May 14, 2007 - link

    No they dont and that's why the 2600 and 2400 don't exist

  • ochentay4 - Monday, May 14, 2007 - link

    Let me start with this: i always had a nvidia card. ALWAYS.

    Faster is NOT ALWAYS better. For the most part this is true, for me, it was. One year ago I boght a MSI7600GT. Seemed the best bang for the buck. Since I bought it, I had problems with TVout detection, TVout wrong aspect ratios, broken LCD scaling, lot of game problems, inexistent support (nv forum is a joke) and UNIFIED DRIVER ARQUITECTURE. What a terrible lie! The latest official drivers is 6 months ago!!!

    Im really demanding, but i payed enough to demand a 100% working product. Now ATi latest offering has: AVIVO, FULL VIDEO ACC, MONTHLY DRIVER UPDATES, ALL BUGS I NOTICED WITH NVIDIA CARD FIXED, HDMI AND PRICE. I prefer that than a simple product, specially for the money they cost!

    I will never buy a nvidia card again. I'm definitely looking forward ATis offering (after the joke that is/was 8600GT/GTS).

    Enough rant.
    Am I wrong?

  • Roy2001 - Tuesday, May 15, 2007 - link

    Yeah, you are wrong. Spend $400 on a 2900XT and then $150 on a PSU.
Copyright © 2024. All rights reserved.
BENCH
TOPICS
FOLLOW
ABOUT

Log in

Don't have an account? Sign up now