1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Crysis / Crysis warhead faster in DX10 on ATi 5k series

Discussion in 'AMD / ATI' started by crazyeyesreaper, Feb 24, 2010.

  1. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    Alright from what ive tested Crysis runs faster in DX10 mode then it does DX9. These are some of my logs from my crysis runs. I plan to do further testing, i want community feed back and we will try and see if its a general performance difference or if im just magically lucky and have pixie dust working magic in my system. Now on to the show

    using: Crysis Benchmark Tool avaible at Guru 3d

    here are my results

    DX9

    Run #1- DX9 1900x1200 AA=2x, 32 bit test, Quality: High ~~ Last Average FPS: 39.56


    Run #1- DX9 1900x1200 AA=2x, 32 bit test, Quality: High ~~ Overall Average FPS: 43.895


    Run #1- DX9 1900x1200 AA=2x, 32 bit test, Quality: High ~~ Overall Average FPS: 44.505


    DX10


    Run #1- DX10 1900x1200 AA=2x, 32 bit test, Quality: High ~~ Last Average FPS: 46.65


    Run #1- DX10 1900x1200 AA=2x, 32 bit test, Quality: High ~~ Overall Average FPS: 47.74


    Run #1- DX10 1900x1200 AA=8x, 32 bit test, Quality: High ~~ Overall Average FPS: 48.47


    Now for parity these were run at default HIGH settings 2xAA 1900x1200 as indicated
    now these are crossfire results ive noticed single card configs get the same small boost ill post that info soon on my own findings.

    [​IMG]
     
  2. CJCerny

    CJCerny

    Joined:
    Nov 25, 2008
    Messages:
    862 (0.38/day)
    Thanks Received:
    253
    Location:
    Akron, OH
    Did you reboot between test runs or did you just click on the DX10 radio button after you were done with the DX9 tests?
     
  3. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    exited out and restarted the app also cleared the runs and reset the settings each time as well

    also dx9 vs dx 10 is easy to spot when watched for and i like to watch my benchmark runs :toast:

    i can confirm having a friend test his 9600gt that for him DX9 was faster then DX10 so im looking for more ati 5770 5850 and 5870 users to post there results ;) for comparisons

    i havent tested warhead yet but considering same engine etc just the draw distance is scaled back in warhead a bit the results should be similar id prefer old crysis comparisons but warhead results are welcome as well

    and for those intrested im on the 9.12 hotfix drivers

    im also aware that it says 8xAA in the last run but notice the oddity the frame rate average went up o_O
     
    Last edited: Feb 24, 2010
  4. bpgt64

    bpgt64

    Joined:
    Oct 5, 2008
    Messages:
    1,524 (0.66/day)
    Thanks Received:
    206
    Location:
    ATL, GA
    Look at those tests, see how it increases in FPS the more AA you turn on...thats awesome
     
  5. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    yea rather comical but then again my Phenom II at only 3.4 is hold these gpus back so i guess giving them more stuff to crunch got them psyched up or something :roll:
     
  6. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    the only thing is with crysis warhead its 1920x1200, is that cool? I remember just crysis bench is 1900x1200.... weird.
     
  7. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    yea thats fine man just as long as DX9 vs DX10 settings are the same ;)
     
  8. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    cool ima run some dx9 vs dx10 warhead tests and post them shortly
     
  9. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    thanks wolf

    im wondering if it has anything to do with how the gpus are more focused toward DX11 and with DX10 being a subset of DX11......
     
  10. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    well I'm pretty sure my DX10 scores are lower than DX9, maybe an i7 @ 4.2ghz lets the GPU's do whatever they need to.. I'm not sure... results soon.

    EDIT: Bah it has to wait a lil, my bench tool uis acting up and not giving me scores after the benches and its 1am and I work tomorrow :)
     
    Last edited: Feb 24, 2010
  11. pantherx12

    pantherx12 New Member

    Joined:
    Jan 2, 2009
    Messages:
    9,714 (4.38/day)
    Thanks Received:
    1,699
    Location:
    ENGLAND-LAND-LAND
    Makes sense, they would make optimisations for more recent things rather then older things :]


    Good stuff though :p

    Cheers for sharing.
     
  12. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    true panter but its still rather intresting go from dx 9 to dx 10 see some improvement go from 2xaa to 8xAA see more improvement its a double take senario i hardly believe it but i can run the bench all day and that numbers dont lie :roll: rather comical that we have reached the DX10 is faster then DX9 in crysis phase lol
     
  13. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    Some Warhead benches, 1920x1200, 4xAA and no AA, DX9 and 10.

    DirectX 10 ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 4x
    ==> Framerate [ Min: 35.13 Max: 68.90 Avg: 52.48 ]

    DirectX 10 ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 0x
    ==> Framerate [ Min: 45.72 Max: 78.50 Avg: 63.96 ]

    DirectX 9.0c ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 4x
    ==> Framerate [ Min: 52.34 Max: 80.62 Avg: 67.20 ]

    DirectX 9.0c ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 0x
    ==> Framerate [ Min: 52.23 Max: 100.46 Avg: 79.84 ]

    she's definitely faster in DX9 mode.

    [​IMG]
     
  14. SilverSRT4turbo

    SilverSRT4turbo New Member

    Joined:
    Feb 6, 2010
    Messages:
    22 (0.01/day)
    Thanks Received:
    2
    Location:
    Tampa, FL
    Anyone have a link to the benchmark program? I have a 5870 (965 BE) all stock and have been running 1900x1080 with natural mod+very high settings and its run great so far (in my eyes atleast) but im sure Ill have to back it down eventually when the action get crazier in the later levels.

    But Ill post my results if that helps you guys out.
     
  15. janolle New Member

    Joined:
    May 17, 2009
    Messages:
    5 (0.00/day)
    Thanks Received:
    0
    Yup, attached my testresults...
     

    Attached Files:

  16. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    the reason i asked for High mode was because DX9 cant really use all the features of enthusiast the highest settings as in game on crysis very high aka enthusiast is locked sure i can choose enthusiast for DX9 in the benchmark app but its not using enthusiast settings but then again im only testing Crysis not warhead as warhead runs like crap for me
     
  17. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    so should I test crysis warheads "gamer" preset too?
     
  18. pantherx12

    pantherx12 New Member

    Joined:
    Jan 2, 2009
    Messages:
    9,714 (4.38/day)
    Thanks Received:
    1,699
    Location:
    ENGLAND-LAND-LAND
    Wolf could you try the same as crazyeyes and max out AA see if there's a hit or increase to fps please?
     
  19. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    I would but I cant get crysis to not crash on me while loading up, weeks ago i was doing some testing with my single 5870 and cat 10.1, but during that testing it bugged out with "crysis.exe has stopped working" and has been doing it ever since.

    I've tried everything but re-downloading all the files from steam...
     
  20. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    DirectX 9.0c GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 2x
    ==> Framerate [ Min: 1.94 Max: 47.10 Avg: 36.61 ]

    DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 2x
    ==> Framerate [ Min: 11.37 Max: 43.52 Avg: 33.95 ]

    these are my crysis warhead results obvious crysis warhead dosent like my system and runs far worse then crysis the original
     
  21. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    I'll mirror your test results right now, in crossfire mode.
     
  22. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    that was crossfire wolf

    warhead runs like absolute crap for me

    crossfire

    crysis min 22 average 44 in dX10

    warhead min 11 average 33 in dx10
     
  23. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    I can see that, those MIN FPS readings are in the toilet really..... and DX10 MIN FPS shits on DX9 still, odd.
     
  24. crazyeyesreaper

    crazyeyesreaper Chief Broken Rig

    Joined:
    Mar 25, 2009
    Messages:
    8,197 (3.84/day)
    Thanks Received:
    2,834
    Location:
    04578
    yup crysis in general shits all over warhead for me which is also odd since everyone touts warhead as the better optimized title
     
  25. wolf

    wolf Performance Enthusiast

    Joined:
    May 7, 2007
    Messages:
    5,547 (1.96/day)
    Thanks Received:
    847
    I am certain crysis (original) runs better than warhead on mine too, but warhead still runs just fine, if only it worked so I could test it.
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page