Closed Solved

Poor 6850 Crossfire Performance! Help!

So I just ran 3DMark Vantage on the default (performance) presets and got a score of only 20000!

Numerous reviews and benchmarks state I should be getting ~ 25k.

What's going on!???

Hell, 2 x 5770's in crossfire get 20,000.

Something is seriously wrong.
If anyone has any advice that would be greatly appreciated. Drivers are up to date.

-----

i5 750 – 3.2 GHz (No bottlenecking here)
4GB 1600Mhz

I should add that one card is an x16 2.0 slot and the other in a x4 2.0 slot.
But for those that don't know it makes almost no difference.
17 answers Last reply Best Answer
More about poor 6850 crossfire performance help
  1. Two 5770's don't normally get to 20,000.
  2. For a single card: http://www.tomshardware.com/reviews/radeon-hd-6870-radeon-hd-6850-barts,2776-9.html

    Are you giving me the full score, the one with the P in front of it, or the GPU score.

    While the GPU score may hit 25k, the full score is going to be closer to 20k
  3. na just the GPU score
  4. with two hd5850's cf I get a gpu score of 24779
  5. Post a GPU-Z screenshot of both GPU's, if you would.

    Oh, BTW, a lot of these Benchmarks are posting with higher OC'ed i5's. While it might not bottleneck most games, it still increases your score and some games still improve the FPS with the CPU OC'ed all the way to 4.0.
  6. I oc'd My i5 750 to 4.0GHz and there was no difference (~10 points - fluctuation) compared with 3.2GHz.

    Which tab of GPU-Z? Which Graphics card? etc.
  7. cammo2066, it does make a slight difference as x4 is slower by quite a lot.
    x16 and x8 have almost no difference but x16 and x4 do.
  8. I was looking at benchmarks, and checking out their CPU clock, and the ones with 25k GPU scores had over 4.0 OC'ed i7's.

    Check this out: http://www.tomshardware.com/reviews/radeon-hd-5850,2433-4.html

    Granted, the 6850 is not on it, but that is with an OC of 3.33. With that OC, the crossfired 5870's only get 21k, which score higher than 6850's.
  9. Not when the x4 is PCI-E 2.0 standard

    http://www.tomshardware.com/reviews/radeon-hd-5850,2433-4.html - Going by this review, crossfired 6850's get a score of 18k which means that jijoslin is a complete liar.

    But I highly doubt that, so why are those benchmarks different?
  10. cammo2066 said:
    Not when the x4 is PCI-E 2.0 standard

    http://www.tomshardware.com/reviews/radeon-hd-5850,2433-4.html - Going by this review, crossfired 6850's get a score of 18k which means that jijoslin is a complete liar.

    But I highly doubt that, so why are those benchmarks different?


    Two major differences. The benchmarks I saw that hit 25k has an i7 OC'ed to 4.2Ghz. Say what you will about your i5 not bottlenecking all you want, but CPU speeds makes a difference and makes a huge difference when you are at 1280x768, which is resolution used on the performance setting.

    The other difference is newer drivers have improved performance some.
  11. cammo2066 said:
    I oc'd My i5 750 to 4.0GHz and there was no difference (~10 points - fluctuation) compared with 3.2GHz.

    Which tab of GPU-Z? Which Graphics card? etc.


    The tab at the bottom left that should have a drop down for both cards. Make sure the bus interface is running at x16 2.0, or at least x8 2.0. Most cases x4 2.0 will kill performance by as much as 40%.
  12. The second slot is x4 2.0; As stated in the O.P.
    As also stated it makes little-no difference. Look here:

    http://www.hardocp.com/article/2010/08/25/gtx_480_sli_pcie_bandwidth_perf_x16x16_vs_x4x4

    With all the information a GTX 480 is pushing through an x4 2.0 Slot there is still no real performance impact at the largest resolution.
  13. Best answer
    cammo2066 said:
    The second slot is x4 2.0; As stated in the O.P.
    As also stated it makes little-no difference. Look here:

    http://www.hardocp.com/article/2010/08/25/gtx_480_sli_pcie_bandwidth_perf_x16x16_vs_x4x4

    With all the information a GTX 480 is pushing through an x4 2.0 Slot there is still no real performance impact at the largest resolution.


    Those are 480's, and they may have tried it on a different board than yours, one with some improvements, but check this one out, with 5870's, which should be closer, as both are ATI cards: http://www.tomshardware.com/reviews/p55-pci-express-scaling,2517-6.html

    That test was also done on an i7 system, which the x4 slots scale better. The one I posted shows both the i7 and i5 (the slower x4 slot was an i5).
  14. How about you toss that software out of the way for just a bit, and make a real benchmark like say, the free Stalker clear sky benchmark or others...

    ppl get weird scores with 3dmark all the time, what matters are the frames per second you get in an actual game.

    If you choose to go for the Stalker benchy, I have some screenshot of several benchmarks with different cards, I just need to locate them first.
  15. bystander said:
    Those are 480's, and they may have tried it on a different board than yours, one with some improvements, but check this one out, with 5870's, which should be closer, as both are ATI cards: http://www.tomshardware.com/reviews/p55-pci-express-scaling,2517-6.html

    That test was also done on an i7 system, which the x4 slots scale better. The one I posted shows both the i7 and i5 (the slower x4 slot was an i5).


    Hmm... I see...

    I guess that's why I'm getting shitty 3DMark Vantage scores.

    However Crysis still runs silky smooth and benchmarks at 3 FPS less than the Guru 3d 6850 crossfire bench does.

    Wierd...

    Oh well, thanks a lot for the information
  16. Best answer selected by cammo2066.
  17. This topic has been closed by Mousemonkey
Ask a new question

Read More

Radeon Performance Crossfire Graphics