Compute Performance

For our look at compute performance this is going to be a brief look. Our OpenGL AES and DirectCompute Fluid Simulation benchmarks simply don’t scale with multiple GPUs, so we’ll skip though (though the data is still available in Bench).

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.

Given the nature of the benchmark, it’s not surprising that we see a performance regression here with some setups. The nature of this benchmark is that it doesn’t split across multiple GPUs well, though that doesn’t stop AMD and NVIDIA from tying. This doesn’t impact real game performance as we’ve seen, but it’s a good reminder of the potential pitfalls of multi-GPU configurations. Though AMD does deserve some credit here for gaining on their single GPU performance, pushing their lead even higher.

Our other compute 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.

Unlike the Civ V compute benchmark, SLG scales very well with multiple GPUs, nearly doubling in performance. Unfortunately for NVIDIA GK104 shows its colors here as a compute-weak GPU, and even with two of them we’re nowhere close to one 7970, let alone the monster that is two. If you’re looking at doing serious GPGPU compute work, you should be looking at Fermi, Tahiti, or the future Big Kepler.

Civilization V Power, Temperature, & Noise
Comments Locked

200 Comments

View All Comments

  • Death666Angel - Thursday, May 3, 2012 - link

    The few reviews I've seen have 4GB GTX 680 card between 5% and 10% faster at high resolutions (starting at 2560x1440 to 7860x1600). Adding, on top of that some more memory bandwidth would have been the gaming card most people expected from nVidia.
    As it stands, the GTX 680 is good, but also very expensive (I can have t he 7970 for 65€ less). The GTX 690 is a good product for people who want SLI but don't have the space, PSU, SLI enabled mainboard or want 4 GPUs.
  • CeriseCogburn - Saturday, May 5, 2012 - link

    Sure, link us to a single review that shows us that. Won't be HardOcp nor any as popular, as every review has shown the exact opposite.
  • kallogan - Thursday, May 3, 2012 - link

    Where are the middle range gpus ?

    Wait. Nvidia don't release them cause they can't provide enough quantities.
  • CeriseCogburn - Thursday, May 3, 2012 - link

    They're being held back like the "real 680" top nVidia core, because nVidia is making more money selling the prior launches and the new 2nd tier now top dog cards.
    It's a conspiracy of profit and win.
  • silverblue - Thursday, May 3, 2012 - link

    Yes, because making a small number of full size Kepler cores is obviously going to make them more money than a large number of less complex Kepler cores. *rolleyes*

    NVIDIA, assuming they had the ability to get them manufactured in large enough quantities, would make far more profit off a 660 or 670 than they ever would off a 680.
  • silverblue - Thursday, May 3, 2012 - link

    (I mean making far more profit off the 660/670 series than the 680 series, not specific cards nor the profit per card)
  • CeriseCogburn - Tuesday, May 8, 2012 - link

    A lot of prior gen stock moving, take a look you're on the internet, not that hard to do, wonder why you people are always so clueless.
  • CeriseCogburn - Tuesday, May 8, 2012 - link

    For instance the entire lot of 7870's and 7850's on the egg are outsold by a single GTX680 by EVGA - confirmed purchasers reviews.
    So it appears nVidia knows what it's doing in a greatly superior manner than your tiny mind spewing whatever comes to it in a few moments of raging rebuttal whilst you try to "point out what's obvious" - yet is incorrect.
  • zcat - Thursday, May 3, 2012 - link

    Ditto.

    Every time my Anandtech feed updates, the first thing I'm hoping to see is reviews for the more-reasonably priced, and less power-hoggy GTX 640 (w/GDDR5) and GTX 660 Ti. If we see a review, then at least we know it'll show up at retail very soon after.

    All I want for xmas is a mid-range NVidia card with a higher idle wattage to maximum performance ration than AMD (because NVidia > AMD wrt drivers, esp under linux).
  • zcat - Thursday, May 3, 2012 - link

    correction: idle Watts <= 10 && max performance >= AMD.

Log in

Don't have an account? Sign up now