Compute Performance

Moving on from our look at gaming performance, we have our customary look at compute performance. With GCN AMD significantly overhauled their architecture in order to improve compute performance, as their long-run initiatives rely on GPU compute performance becoming far more important than it is today.

Our first compute benchmark comes from Civilization V, which uses DirectCompute to decompress textures on the fly. Civ V includes a sub-benchmark that exclusively tests the speed of their texture decompression algorithm by repeatedly decompressing the textures required for one of the game’s leader scenes. Note that this is a DX11 DirectCompute benchmark.

The Civ V compute shader benchmark once again shows off just how much the compute shader performance of the 7800 series has improved relative to the 6900 series, with both 7800 cards coming in well, well ahead of any previous generation AMD cards. Compared to NVIDIA’s lineup the 7800 series does fairly well for itself too, although not quite as well as the commanding lead the 7900 series took.

Our next benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. We’re now using a development build from the version 2.0 branch, and we’ve moved on to a more complex scene that hopefully will provide a greater challenge to our GPUs.

SmallLuxGPU continues to showcase the 7800 series’ improvements over past AMD architectures, and while it’s not the same kind of massive leap we saw with CivV, it’s still enough to bring the 7850 up to near the performance of the 6970, and pushing the 7870 well beyond that. The only real competition here for AMD is AMD.

For our next benchmark we’re looking at AESEncryptDecrypt, an OpenCL AES encryption routine that AES encrypts/decrypts an 8K x 8K pixel square image file. The results of this benchmark are the average time to encrypt the image over a number of iterations of the AES cypher.

On the one hand, the 7870 gets quite close to the 7950 here in our AESEncryptDecrypt benchmark, in spite of the latter’s higher number of shaders. On the other hand, it’s still not enough to dethrone the GTX 570; the only NVIDIA cards the 7800 series can beat start at the GTX 560 Ti.

Finally, our last benchmark is once again looking at compute shader performance, this time through the Fluid simulation sample in the DirectX SDK. This program simulates the motion and interactions of a 16k particle fluid using a compute shader, with a choice of several different algorithms. In this case we’re using an (O)n^2 nearest neighbor method that is optimized by using shared memory to cache data.

In our final compute test the 7800 series once again makes a run at the top, with both cards rising past the GTX 570, although they can’t quite match the GTX 580. In an interesting turn of events the 7870 ends up being some 6% faster than the 7950, in spite of the fact that in a compute benchmark the 7950 should have a solid lead. This just goes to show that core clockspeeds do matter, and that adding more shaders alone can’t conquer all benchmarks.

Civilization V Theoretical Performance
Comments Locked

173 Comments

View All Comments

  • CeriseCogburn - Thursday, March 8, 2012 - link

    Yet if one is not running gigantic resolutions, they look at the usual, 1920 and 1650 p resolutions, and likely want to crank all the eye candy to the limit, which is still IMPOSSIBLE at those common resolutions with 60+ frames in so many of the popular games.
    So the real problem is you go from "can't do it all" to "still can't do it all" but at least you've got 40 frames going to 55 on your one screen... with maybe one more setting of 7 at ultra...
    ---
    For others with 3x 2560 most of us really don't give a crap if they claim they get 2x frames - because if they don't have 2 or 3 or 4 of them running, they are stuck in turn down the eye candy crapsville TOO.
    --
    We almost always hear that we are stuck with console ports, the exact opposite of the real truth in the real problem.
    A 570 is NOT ENOUGH, a 7870 is not either, nor is a 7970 for 1900x1200.
    IT'S NOT GOOD ENOUGH.
  • SlyNine - Thursday, March 8, 2012 - link

    75% was wrong, but so is 70-110% faster.

    Crysis 2560x1600 at the MAX settings its 20 vx 33, which is just over 60%

    Drop down to 1920x1200 at the advantage drops to 50%.

    Metro its 60% at max res/settings (36vs22.5)

    Drop that down to 1920X1200 and its just over 50%

    Dirt 3 its just over 50% at max res/settings Drop that down to 1920x1200 and its remains just above 50% ( 104 vs 68.4)

    Battlefield 3 its 50% (49.7 vs 32.6) at max settings/res.

    So where the heck are you getting 70% to 110% ??

    2 1/2 years ago I payed around 66% of your price, and I'm getting 66% of your performence, ALMOST 3 YEARS. THIS IS NOT MOVING THINGS FORWARD
  • CeriseCogburn - Thursday, March 8, 2012 - link

    Get back to us krummer when the 7870 is "released" and has "stable drivers" that "work most of the time" in "most of the games" and the IQ cheats of 10% driver default plus ever more now with this new blur job called MLAA and the lack of LOD bias up high enough that "in the case of SSAA" it's another low detail IQ cut down, not to mention other things like PhysX and tessellation above 10 all the way to 32...the other "unneeded" "eye candy" that "sucks" because amd sucks at doing it.
    *
    9% , minus 10% standardized cheat, minus SSAA LOD bias cheat, minus MLAA blur cheat, minus PhysX, minus 8 other things I won't take time to mention doesn't sound like "faster" to me.
    I mean come on, if the arch is so superior, why all the hack and cheats and crappy blurring and lack of features ?
    Are the drivers going to be another ongoing nightmare for 47% of all ibm pc implementations ?
  • chizow - Monday, March 5, 2012 - link

    @ Kiste: Agreed.

    The "competition is necessary" meme needs to die in the tech sector because it isn't necessary.

    Most of this stuff doesn't expire or die on its own, not before it becomes obsolete anyways and in order for it to become obsolete there needs to be innovation and performance increases.

    That's what drives innovation with technology and it certainly exists without competition.
  • medi01 - Monday, March 5, 2012 - link

    What are you talking about, considering AMD 7850 is faster in most tests than nVidia's 80$ more expensive 570, at the same time consuming 25% less energy.
  • Malih - Monday, March 5, 2012 - link

    I think this is the decision of the new management, they decide to price something that performs better to price higher.

    Probably will drop the price in the future, but it would require a new release from nVidia with agressive pricing. It is rumored nVidia will release new cards near the end of March.
  • biassj - Monday, March 5, 2012 - link

    Shitty pricing again, if the pricing was 50 bucks cheaper I would probably consider buying 7870 or 7950 at this moment. These high prices will just have me wait to see what Nvidia has to offer.
  • piroroadkill - Monday, March 5, 2012 - link

    Huh, I thought it was GTX 570 class for less.

    Seems solid enough.
  • piroroadkill - Monday, March 5, 2012 - link

    For the 7850, I mean.
  • chizow - Monday, March 5, 2012 - link

    This pricing isn't nearly as bad as the 7770 or 7950/7970, but its still pretty poor overall given it once again, slides right in to existing price structures offering very little incentive to upgrade and very little price performance value compared to what has been available for 14+ months.

Log in

Don't have an account? Sign up now