AA Image Quality & Performance

With HL2 unsuitable for use in assessing image quality, we will be using Crysis: Warhead for the task. Warhead has a great deal of foliage in parts of the game which creates an immense amount of aliasing, and along with the geometry of local objects forms a good test for anti-aliasing quality. Look in particular at the leaves both to the left and through the windshield, along with aliasing along the frame, windows, and mirror of the vehicle. We’d also like to note that since AMD’s SSAA modes do not work in DX10, this is done in DX9 mode instead.


AMD Radeon HD 5870 - No AA

AMD Radeon HD 5870
AMD Radeon HD 4870
NVIDIA GTX 280
No AA
2X MSAA
4X MSAA
8X MSAA
2X MSAA +AAA 2X MSAA +AAA 2X MSAA + SSTr
4X MSAA +AAA 4X MSAA +AAA 4X MSAA + SSTr
8X MSAA +AAA 8X MSAA +AAA 8X MSAA + SSTr
2X SSAA    
4X SSAA    
8X SSAA    


From an image quality perspective, very little has changed for AMD compared to the 4890. With MSAA and AAA modes enabled the quality is virtually identical. And while things are not identical when flipping between vendors (for whatever reason the sky brightness differs), the resulting image quality is still basically the same.

For AMD, the downside to this IQ test is that SSAA fails to break away from MSAA + AAA. We’ve previously established that SSAA is a superior (albeit brute force) method of anti-aliasing, but we have been unable to find any scene in any game that succinctly proves it. Shader aliasing should be the biggest difference, but in practice we can’t find any such aliasing in a DX9 game that would be obvious. Nor is Crysis Warhead benefitting from the extra texture sampling here.

From our testing, we’re left with the impression that for a MSAA + AAA (or MSAA + SSTr for NVIDIA) is just as good as SSAA for all practical purposes. Much as with the anisotropic filtering situation we know through technological proof that there is better method, but it just isn’t making a noticeable difference here. If nothing else this is good from a performance standpoint, as MSAA + AAA is not nearly as hard on performance as outright SSAA is. Perhaps SSAA is better suited for older games, particularly those locked at lower resolutions?

For our performance data, we have two cases. We will first look at HL2 on only the 5870, which we ran before realizing the quality problem with Source-engine games. We believe that the performance data is still correct in spite of the visual bug, and while we’re not going to use it as our only data, we will use it as an example of AA performance in an older title.

As a testament to the rendering power of the 5870, even at 2560x1600 and 8x SSAA, we still get a just-playable framerate on HL2. To put things in perspective, with 8x SSAA the game is being rendered at approximately 32MP, well over the size of even the largest possible single-card Eyefinity display.

Our second, larger performance test is Crysis: Warhead. Here we are testing the game on DX9 mode again at a resolution of 1920x1200. Since this is a look at the impact of AA on various architectures, we will limit this test to the 5870, the GTX 280, and the Radeon HD 4890. Our interest here is in performance relative to no anti-aliasing, and whether different architectures lose the same amount of performance or not.


Click to Enlarge

Starting with the 5870, moving from 0x AA to 4x MSAA only incurs a 20% drop in performance, while 8x MSAA increases that drop to 35%, or 80% of the 4x MSAA performance. Interestingly, in spite of the heavy foliage in the scene, Adaptive AA has virtually no performance hit over regular MSAA, coming in at virtually the same results. SSAA is of course the big loser here, quickly dropping to unplayable levels. As we discussed earlier, the quality of SSAA is no better than MSAA + AAA here.

Moving on, we have the 4890. While the overall performance is lower, interestingly enough the drop in performance from MSAA is not quite as much, at only 17% for 4x MSAA and 25% for 8x MSAA. This makes the performance of 8x MSAA relative to 4x MSAA 92%. Once again the performance hit from enabling AAA is miniscule, at roughly 1 FPS.

Finally we have the GTX 280. The drop in performance here is in line with that of the 5870; 20% for 4x MSAA, 36% for 8x MSAA, with 8x MSAA offering 80% of the performance. Even enabling supersample transparency AA only knocks off 1 FPS, just like AAA under the 5870.

What this leaves us with are very curious results. On a percentage basis the 5870 is no better than the GTX 280, which isn’t an irrational thing to see, but it does worse than the 4890. At this point we don’t have a good explanation for the difference; perhaps it’s a product of early drivers or the early BIOS? It’s something that we’ll need to investigate at a later date.

Wrapping things up, as we discussed earlier AMD has been pitching the idea of better 8x MSAA performance in the 5870 compared to the 4800 series due to the extra cache. Although from a practical perspective we’re not sold on the idea that 8x MSAA is a big enough improvement to justify any performance hit, we can put to rest the idea that the 5870 is any better at 8x MSAA than prior cards. At least in Crysis: Warhead, we’re not seeing it.

The Return of Supersample AA The Test
Comments Locked

327 Comments

View All Comments

  • mapesdhs - Saturday, September 26, 2009 - link


    MODel3 writes:
    > 1.Geometry/vertex performance issues ...
    > 2.Geometry/vertex shading performance issues ...

    Would perhaps some of the subtests in 3DMark06 be able to test this?
    (not sure about Vantage, never used that yet) Though given what Jarred
    said about the bandwidth and other differences, I suppose it's possible
    to observe large differences in synthetic tests which are not the real
    cause of a performance disparity.

    The trouble with heavy GE tests is, they often end up loading the fill
    rates anyway. I've run into this problem with the SGI tests I've done
    over the years:

    http://www.sgidepot.co.uk/sgi.html">http://www.sgidepot.co.uk/sgi.html

    The larger landscape models used in the Inventor tests are a good
    example. The points models worked better in this regard for testing
    GE speed (stars3/star4), but I don't know to what extent modern PC
    gfx is designed to handle points modelling - probably works better
    on pro cards. Actually, Inventor wasn't a good choice anyway as it's
    badly CPU-bound and API-heavy (I should have used Performer, gives
    results 5 to 10X faster).

    Anyway, point is, synthetic tests might allow one to infer that one
    aspect of the gfx pipeline is a bottleneck when infact it isn't.

    Ages ago I emailed NVIDIA (Ujesh, who I used to know many moons ago,
    but alas he didn't reply) asking when, if ever, they would add
    performance counters and other feedback monitors to their gfx
    products so that applications could tell what was going on in the
    gfx pipeline. SGI did this ages years ago, which allowed systems like
    IR to support impressive functions such as Dynamic Video Resizing by
    being able to monitor frame by frame what was going on within the gfx
    engine at each stage. Try loading any 3D model into perfly, press F1
    and click on 'Gfx' in the panel (Linux systems can run Performer), eg.:

    http://www.sgidepot.co.uk/misc/perfly.gif">http://www.sgidepot.co.uk/misc/perfly.gif

    Given how complex modern PC gfx has become, it's always been a
    mystery to me why such functions haven't been included long ago.
    Indeed, for all that Crysis looks amazing, I was never that keen on
    it being used as a benchmark since there was no way of knowing
    whether the performance hammering it created was due to a genuinely
    complex environment or just an inefficient gfx engine. There's still
    no way to be sure.

    If we knew what was happening inside the gfx system, we could easily
    work out why performance differences for different apps/games crop
    up the way they do. And I would have thought that feedback monitors
    within the gfx pipe would be even more useful to those using
    professional applications, just as it was for coders working on SGI
    hardware in years past.

    Come to think of it, how do NVIDIA/ATI even design these things
    without being able to monitor what's going on? Jarred, have you ever
    asked either company about this?

    Ian.

  • JarredWalton - Saturday, September 26, 2009 - link

    I haven't personally, since I'm not really the GPU reviewer here. I'd assume most of their design comes from modeling what's happening, and with knowledge of their architecture they probably have utilities that help them debug stuff and figure out where stalls and bottlenecks are occurring. Or maybe they don't? I figure we don't really have this sort of detail for CPUs either, because we have tools that know the pipeline and architecture and they can model how the software performs without any hardware feedback.
  • MODEL3 - Thursday, October 1, 2009 - link

    I checked the web for synthetic geometry tests.
    Sadly i only found 3dMark Vantage tests.
    You can't tell much from them, but they are indicative.

    Check:

    http://www.pcper.com/article.php?aid=783&type=...">http://www.pcper.com/article.php?aid=783&type=...

    GPU Cloth: 5870 is only 1,2X faster than 4890. (vertex/geometry shading test)
    GPU Particles: 5870 is only 1,2X faster than 4890. (vertex/geometry shading test)

    Perlin Noise: 5870 is 2,5X faster than 4890. (Math-heavy Pixel Shader test)
    Parallax Occlusion Mapping: 5870 is 2,1X faster than 4890. (Complex Pixel Shader test)

    All the above 4 tests are not bandwidth limited at all.
    Just for example, if you check:

    http://www.pcper.com/article.php?aid=674&type=...">http://www.pcper.com/article.php?aid=674&type=...

    You will see that a 750MHz 4870 512MB is 20-23% faster than a 625MHz 4850 in all the above 4 tests, so the extra bandwidth (115,2GB/s vs 64GB/s) it doesn't help at all.
    But 4850 is extremely bandwidth limited in the color fillrate test (4870 is 60% faster than 4850)

    Also it shouldn't be a problem of the dual rasterizer/dual SIMDs engine efficiency since synthetic Pixel Shader tests is fine (more than 2X) while the synthetic geometry shading tests is only 1,2X.

    My guess is ATI didn't improve the classic geometry set-up engine and the GS because they want to promote vertex/geometry techniques based on the DX11 tesselator from now on.
  • Zool - Friday, September 25, 2009 - link

    In Dx11 the fixed tesselation units will do much finer geometry details for much less memmory space and on chip so i think there isnt a single problem with that. Also the compute shader need minimal memory bandwith and can utilize plenty of idle shaders. The card is designed with dx11 in mind and it isnt using the wholle pipeline after all. I wouldnt make too early conclusions.(I think the perfomance will be much better after few drivers)

  • MODEL3 - Saturday, September 26, 2009 - link

    The DX11 tesselator in order to be utilized must the game engine to take advantage of it.
    I am not talking about the tesselator.
    I am talking about the classic Geometry unit (DX9/DX10 engines) and the Geometry Shader [GS] (DX10 engines only).

    I'll check to see if i can find a tech site that has synthetic bench for Geometry related perf. and i will post again tomorrow, if i can find anything.

  • JarredWalton - Friday, September 25, 2009 - link

    It's worth noting that when you factor in clock speeds, compared to the 5870 the 4870X2 offers 88% of the core performance and 50% more bandwidth. Some algorithms/games require more bandwidth and others need more core performance, but it's usually a combination of the two. The X2 also has CrossFire inefficiencies to deal with.

    More interesting perhaps is that the GTX 295 offers (by my estimates, which admittedly are off in some areas) roughly 10% more GPU shader performance, about 18.5% more fill rate, and 46% more bandwidth than the HD 5870. The fact that the HD 4870 is still competitive is a good sign that ATI is getting good use of their 5 SPs per Stream Processor design, and that they are not memory bandwidth limited -- at least not entirely.
  • SiliconDoc - Wednesday, September 30, 2009 - link

    The 4870x2 has somewhere around "double the data paths" in and out of it's 2 cpu's. So what you have with the 5870 putting as some have characterized " 2x 770 cores melded into one " is DOUBLE THE BOTTLENECK in and out of the core.
    They tried to compensate with ddr5 1200/4800 - but the fact remains, they only get so much with that "NOT ENOUGH DATA PATHS/PINS in and out of that gpu core."
  • cactusdog - Friday, September 25, 2009 - link

    Omg these cards look great. Lol Silicon Doc is so gutted and furious he is making hmself look like a dam fool again only this time he should be on suicide watch...Nvidia cards are now obsolete..LOL.
  • mapesdhs - Friday, September 25, 2009 - link


    Hehe, indeed. Have you ever seen a scifi series called, "They Came
    From Somewhere Else?" S.D.'s getting so worked up, reminds me of
    the scene where the guy's head explodes. :D

    Hmm, that's an alternative approach I suppose in place of post
    moderation. Just get someone so worked up about something they'll
    have an aneurism and pop their clogs... in which case, I'll hand
    it back to Jarred. *grin*

    Ian.

  • SiliconDoc - Friday, September 25, 2009 - link

    That is quite all right, you fellas make sure to read it all, I am more than happy that the truth is sinking into your gourds, you won't be able to shake it.
    I am very happy about it.

Log in

Don't have an account? Sign up now