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
POST A COMMENT

199 Comments

View All Comments

  • james.jwb - Thursday, May 03, 2012 - link

    You are correct, I don't own one... I own three in triple screen. Dell U2412m's.

    I really am at a loss as to what you are on about. It is well known that 16:10 is preferred amongst enthusiasts and professionals for a few reasons. If you want 16:9, fine, go for it, but don't act like it's weird that AT are benching with 16:10 just because you went with cheap ass 16:9 screens.
    Reply
  • CeriseCogburn - Friday, May 04, 2012 - link

    Yes of course you are at a loss, you don't understand a word so why reply ?
    You're all at a loss.
    ROFL
    Reply
  • yelnatsch517 - Friday, May 04, 2012 - link

    Are you being sarcastic or an idiot?
    From my experience 1900x1200 24" monitors are the MAJORITY. My work has roughly 50 24" monitors all in that resolution. My HP ZR24W is 1900x1200 as well. The only 24" monitor that I have even seen is the TN panel that came with an HP computer.

    If you are talking about computer monitors, 1900x1200 is the dominant resolution. If you are talking about TVs, then obviously 1080p is the norm.
    Reply
  • CeriseCogburn - Saturday, May 05, 2012 - link

    There are 242 - count them, well over 200, nearly 250 1920X1080 monitors at the egg.
    _
    In your great experience, there are 16 that fit your 1920X1200 dreampipe FUD scenario at the egg, with most of them, well over half, over $400 each, while the 242 common monitors you all buy as you pinch every penny and whine about $10 difference in videocard prices are well under $200 each a lot of the time.
    So now suddenly, you all spend way over $300 to plus $400 for 11% more pixels... ROFL HAHAHHAHHA instead of $150 or $200...
    I guess that's why this place is so biased, the little bloggers are just as whacked when it comes to being honest.
    Reply
  • InsaneScientist - Saturday, May 05, 2012 - link

    Good grief... resorting to personal attacks isn't exactly a good way to get people to listen to you.

    I'm not going to argue that 1080p isn't more common (from what I've read, no one is), because it is more common, you are quite correct there, however I must point out that your logic to arrive at that conclusion is faulty:
    You're contending that 1080p is more common (it is) because there are more models available on Newegg, but just knowing how many models are available doesn't tell us how many units those move.
    If, for example, each of those 22 models of 1920x1200 monitors moves 10 times as much stock as each of the 1920x1080, nearly as many of the 1920x1200 will have been sold as the 1920x1080 ones.
    Now, I don't think that's likely, and I do agree with you that 1080p is more common nowadays (see next point), but your argument is invalid, even though you have come to the correct conclusion.
    Consider this: there are currently two models of iPhone available, compared to dozens of Android phones. By the same logic as you're using, I could say that the iPhone is incredibly rare - I'd get laughed out of town if I tried to make that argument.

    The second point is that 1920x1200 hasn't been nearly as rare in the past as it is today. When I bought my previous monitor and my laptop (both 1920x1200), 1080p monitors were almost unheard of. Since monitors tend to last a while, it's not at all unreasonable for a disproportionate amount of people to be using them compared to their current sales.

    Thirdly, there is a point of diminishing returns. Notice the complete lack of any benchmarks at or below 1680x1050? These cards are so fast that comparisons at those resolutions are pointless - they're all fast enough for anything you could do to them at that screen res - even Crysis. 1920x1080 almost falls into that category, heck, even 1920x1200 almost falls into that category. Benchmarks are only about who wins if there is some advantage to winning. Below 2560x1600, which card you're using is almost completely irrelevant, so why does it even matter whether they used 1920x1080 or 1920x1200?
    Reply
  • CeriseCogburn - Tuesday, May 08, 2012 - link

    Blah blah blah blah and I'm still 100% correct and you are not at all. Reply
  • Decembermouse - Tuesday, May 08, 2012 - link

    You're quite a character. Reply
  • anirudhs - Thursday, May 03, 2012 - link

    I use 2 at work - HP ZR24W. Reply
  • piroroadkill - Sunday, May 06, 2012 - link

    Hm, odd.
    Not only do I have 1920x1200 monitor on my desktop, I have TWO laptops with 1920x1200 screens. Using one right now.
    Yes, they're rarer than 1080p screens, but this is a site for enthusiasts, therefore, it is more likely.
    Reply
  • Ryan Smith - Thursday, May 03, 2012 - link

    The truth is a bit more simple than that. 5760x1200 is because our choice in monitors for multi-monitor testing was based on my personal monitor, which is another PA246Q. NVIDIA's limited display flexibility (same res + same sync) meant that it was easiest to just pair the PA246Q with some more PA246Qs, Consequently it's easier to just test these monitors at their native resolution when we're using NVIDIA cards. Reply

Log in

Don't have an account? Sign up now