Compute Performance

Moving on from our look at gaming performance, we have our customary look at compute performance. Since compute performance is by definition shader bound, the 7950 is at a bit of a disadvantage here compared to gaming performance. Whereas ROP performance scales with the core clock, shader performance is hit by both the reduction in the core clock and the disabled CU array.

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.

AMD’s greatly improved compute performance continues to shine here, though in the case of Civilization V it’s largely consumed by just closing the previously large gap between the GTX 500 series and the Radeon HD 6000 series. As a result the 7950 falls ever so short of the GTX 580, while the factory overclocked Sapphire and XFX cards give the 7950 enough of a push to come within 5% of the 7970.

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.

Under SmallLuxGPU the 7970 enjoyed a large lead over the GTX 580, and this continues with the 7950. Even though the 7950 is well behind the 7970—to the tune of 24%—it’s still 33% ahead of the GTX 580 and the lead only grows from there. Meanwhile the XFX and Sapphire cards can catch up to the 7970 somewhat, but as this is truly a shader-bound test, you can’t make up for the lack of shaders units on the 7950.

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.

In spite of being a compute benchmark, AESEncryptDecrypt is not particularly sensitive to GPU performance, showcasing the impact that setup times can have. The 7950 trails the 7970 by 10%, and overclocking doesn’t change this much. Unfortunately for AMD NVIDIA is still the leader here, showing that AMD’s compute performance still has room to grow.

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.

With the compute shader fluid simulation we once again shift back into a compute task that’s much more shader-bound. The 7950 only reaches 80% of the performance of the 7970, once more proving the real impact of losing a CU array. This is still enough to handily surpass the GTX 580 however, with the 7950 taking a 15% lead.

Civilization V Power, Temperature, & Noise
POST A COMMENT

259 Comments

View All Comments

  • Galidou - Sunday, February 05, 2012 - link

    Everyone is against you but still after all that, you continue to beleive in whatever nonsense you say from the beginning. Not even one person in here supported your opinions, whatever the market is, it just means one thing, you are a selfish nvidia HARDCORE fanboy who gets paid by the green goblin to speak your nonsense.....

    AMD fanboys are intense at times, but you clearly are a madman, the world would say you're wrong that you'd still beleive you're better than everyone else, but if you were, you'd be rich and not spending your time speaking on forums like this one.
    Reply
  • SlyNine - Sunday, February 05, 2012 - link

    You're kidding right rarson.
    What metric do you think we should use to determine value if not through precedence.

    If you're questioning using precedent as a metering stick for value. Hell even the legal system uses precedents to try and determine law, but to you it's not good enough for video cards.
    Reply
  • SlyNine - Sunday, February 05, 2012 - link

    yea based on old nods. new nods bring cheaper parts with better performence. THATS WHAT THIS IS ALL ABOUT!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Reply
  • SlyNine - Sunday, February 05, 2012 - link

    So the 5870 was priced according to the competition.

    No you're full of shit because new cards always bring better performence for the price.
    Reply
  • SlyNine - Sunday, February 05, 2012 - link

    Actually you fail to use logic plenty of times in your arguments. Also I call your consistency in to question.

    Like for instance, the simple fact that new nods= cheaper performence, not better performence that scales with price.

    Not sure why you cannot get that simple bit of logic figured out.
    Reply
  • Deo Domuique - Wednesday, February 08, 2012 - link

    Hey chizow, your comments kept me eventually away from buying the 7950. I use 12 years gaming PCs and always I had Nvidia. This time would be the first that I'd go to AMD's camp.

    Although, I already had too many concerns ( especially the high price ), I could't keep myself. But finally, I did. I'll wait for Kepler. If the 7950's price currently was 320-330€ and not 410€, right now I'd have already the card.
    Reply
  • Apis - Sunday, April 22, 2012 - link

    They are currently dropping the prices for the 7xxx series, I'm found one sapphire 7950 3gb for 360€ when looking right now. Reply
  • Apis - Sunday, April 22, 2012 - link

    In this "Compute: Civilization V" benchmark Radeon HD 5870 got 154.9 FPS
    In the 6970 review, http://www.anandtech.com/show/4061/amds-radeon-hd-... the 5870 got 181.9 FPS.

    Why the regression?
    Reply
  • Wolfpup - Friday, July 20, 2012 - link

    Traditionally Nvidia's hardware has smashed AMD's for Folding, but it looks like PROBABLY the 7970 and 7950 should be respectable with the new architecture...

    I can't actually find ANYONE talking about the issue though or benchmarking...
    Reply

Log in

Don't have an account? Sign up now