VGA Charts 2008: 101 Configurations Tested

Five Generations Tested

Our largest and most up-to-date 3D comparison covers 101 different graphics card configurations spanning five technology generations. The test includes the Radeon X800, X1000, HD 2000, HD 3000, GeForce 6, 7, 8, 9 families, as well as the new HD 4000 series from AMD and both of the GTX 200 models from Nvidia. A total of 500 hours of work was invested in the 6,767 test values, which consist of tests taken at three resolutions (1280x1024, 1680x1050 and 1920x1200 pixels) both with and without anisotropic filtering (AF) and anti-aliasing (AA).

The local grid provided more than 110,000 watts of power for our tests, while the neighbors were serenaded by the maximum noise level of 58.3 dB(A) when running AMD’s Radeon HD 2900 XT CrossFire and X1950 CrossFire. The GTX 260 SLI and 8800 Ultra SLI reached temperatures of over 100 degrees C, causing the GPUs with no additional fans to give up and either choke thermally or crash the PC.

With an artificially-imposed speed limit set by the Core 2 Extreme X6800 running at 2.93 GHz, the current dual-card solutions made up of two Radeon HD 4870s, GeForce GTX 260s or GTX 280s showed very little or no improvement in overall performance. The SLI and CrossFire configurations are frequently slower than a single card due to badly optimized games. In this test environment, the right benchmark and the selected resolution determine whether purchasing the most powerful 3D boards is actually worthwhile. Performance differences are most perceptible at 1680x1050 or 1920x1200 resolutions when AA is enabled.

Editor’s Note: The following five pages will tell you all about the cards we tested, the benchmarks we ran, the settings we chose, the bugs we ran into, and the results we saw. On the last page of this introduction to our newest charts, right underneath our recommendations on which cards to consider, you’ll find a link straight to the data. Use that page for all of your own product comparisons.

  • chise1
    Thanks guys!
    Reply
  • amdfangirl
    Very good! Thanks guys!

    If only you tested F@H PPD...
    Reply
  • rocky1234
    Not to nit pick but ummm where is the Radeon 4870x2 2GB card in those charts or did I miss it you know the one with 2GB memory umm 800x2 shaders GDDR5 memory 3.6Ghz rated at & 750 core x2 whatever.
    Reply
  • c0d3r
    So totally Awesome :D Gooood Job THG! What wud i do without u !
    Reply
  • Mathos
    Just curious. Why is it you're using catalyst 8.6 for the ATI cards which doesn't have proper support for the 4000 series cards, instead of 8.7 or 8.8? Yet for the Nvidia cards, you use the most up to date drivers.
    Reply
  • cangelini
    Rocky1234Not to nit pick but ummm where is the Radeon 4870x2 2GB card in those charts or did I miss it you know the one with 2GB memory umm 800x2 shaders GDDR5 memory 3.6Ghz rated at & 750 core x2 whatever.
    As mentioned in the charts introduction, these numbers take hours upon hours to compile, which means setting cut-offs for the product submissions and drivers. Unfortunately, the X2 didn't make it. However, there are results for a pair of 4870s in CrossFire, which is a roughly equivalent configuration. You'll also notice that there are no 4600-series Radeons. Again, same issue.
    Reply
  • ossie
    Why do I have the impression that at THG is a persistent confusion between power and energy?
    Reply
  • JeanLuc
    Great job, I'm always pointing people to the charts now they can compare the latest hardware.

    The charts also show the difference between games they are coded well and ones that aren't. Compare Crysis (very high quality) and HL2 Ep2 at 1920 x 1200 4xAA 8xAF, on Crysis even the best hardware left begging for mercy at around 24fps, yet HL2 EP2 which looks just as good (graphically speaking) IMO will run at 30 FPS on a 8600 GTS (yes an 8600 GTS!) and X1800XL. A small portion of the difference can be put down to DX10 but not all of it.
    Reply
  • Wonder why you insist on using catalyst 8.6 - a 3/4 months old driver, out *before* the whole HD4xx0 series and not really supporting it. And by the time the 48x0 cards were released you also had a better driver with actual support the the cards (and some 10% performance boost).
    Reply
  • buzzlightbeer
    OMFG they did it again, why oh why did they use 8.6 drivers didnt they learn from all the annoyed people from the last test
    Reply