The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.

    M17x R2 - 3dmark 11 scores on single 6970m/6990m

    Discussion in 'Alienware 17 and M17x' started by flingin, Jan 23, 2012.

  1. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    Hi all,
    i was looking at futuremark web site, and searched for scores with single 6970m/6990m card.

    My score is about 3400, with slight oc becouse of undervolted GPU.

    how this is possible that single (propably oc'd -hell oc'd) 6990m, with 2920xm is 70% faster (5700) than 6970m with OCd 920xm ?
     
    flingin, Jan 23, 2012
    #1
  2. YodaGoneMad

    YodaGoneMad Notebook Deity

    Reputations:
    555
    Messages:
    1,382
    Likes Received:
    12
    Trophy Points:
    56
    My guess is they turned tessellation off and cheated to get that score. Also, it is silly to compare whole 3Dmark scores, you are better to compare just the graphics or just the CPU score, since that will tell you a lot more.
     
    YodaGoneMad, Jan 23, 2012
    #2
  3. long2905

    long2905 Notebook Virtuoso

    Reputations:
    2,443
    Messages:
    2,314
    Likes Received:
    114
    Trophy Points:
    81
    Stock score for 6990M is around 3400-3500 while oc'd is 3900-4200. Turn off tesselation and you got 4-500 more.
     
    long2905, Jan 23, 2012
    #3
  4. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    yes, i did turn tesselation off and i got 3700, still far away from m18x with 5700 :/
     
    flingin, Jan 23, 2012
    #4
  5. long2905

    long2905 Notebook Virtuoso

    Reputations:
    2,443
    Messages:
    2,314
    Likes Received:
    114
    Trophy Points:
    81
    Do you have a single or sli configuration?
     
    long2905, Jan 23, 2012
    #5
  6. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    this are results for single gpu configs only :)
     
    flingin, Jan 23, 2012
    #6
  7. The Revelator

    The Revelator Notebook Prophet

    Reputations:
    5,395
    Messages:
    4,571
    Likes Received:
    2
    Trophy Points:
    105
    @flingin -- It's not possible. For scores 70% higher than your 3400+/-, the scores you're comparing would need be at least 5800+, which means dual 6970M/6990M's in CFX. All you need is a second card and a crossfire cable to even things up.
     
    The Revelator, Jan 23, 2012
    #7
  8. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    flingin, Jan 23, 2012
    #8
  9. long2905

    long2905 Notebook Virtuoso

    Reputations:
    2,443
    Messages:
    2,314
    Likes Received:
    114
    Trophy Points:
    81
    5549 gpu points for a single card? I think there was something glitchy when that person ran 3dmark. Otherwise it would have been the target for many to pursue (since 8/6/2011)
     
    long2905, Jan 23, 2012
    #9
  10. The Revelator

    The Revelator Notebook Prophet

    Reputations:
    5,395
    Messages:
    4,571
    Likes Received:
    2
    Trophy Points:
    105
    That link won't load for me. Will try later.

    If you expect a single undervolted 6970M/920XM to perform like that, you probably ought to go ahead and toss it out that window.

    Without unduly belaboring the point, here is a sample of typical scores for stock 6970M/6990M CFX setups taken from the M18x forum during June, 2011, the date your alleged single card M18x scored P5817.

    Ray - Stock 5817: 2920XM/6970M CFX

    [​IMG]
    http://forum.notebookreview.com/ali...are-m18x-benchmark-thread-33.html#post7561584

    Doc – Stock ~P5817: 2920xm/6970M CFX

    [​IMG]
    See: http://forum.notebookreview.com/ali...are-m18x-benchmark-thread-33.html#post7562286

    Eliphion - Stock ~P5500: 2720qm/6970CFX

    [​IMG]
    http://forum.notebookreview.com/ali...are-m18x-benchmark-thread-41.html#post7601981

    Aikimox - Stock ~P5700: 2630qm/6970CFX

    [​IMG]
    http://forum.notebookreview.com/ali...are-m18x-benchmark-thread-41.html#post7602186

    There are many more examples of the baseline performance of dual 6970M's and 6990M's in CFX. They consistently return about P5500-P6000 (somewhat more with tesselation disabled) in 3dMark11, with CPU's having relatively minor effect on overall scores. A single 6970M or 6990M simply cannot perform at that same level with half the processing power and bandwidth. The 3dMark sheet you refer to does show 1 GPU. It is a 3dMark.com error, which is nothing unusual. It very probably also reports your 920XM as an i7-950, but that does not make it so. You have a nice system. Enjoy it, but don't expect it to perform at levels far beyond its capabilities.
     
    The Revelator, Jan 23, 2012
    #10
  11. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    Well, i know that i can't expect much from single 6970m, the thing i was refering too is that score is huge based on system config.
    I thought that this is single gpu system extreamly overclocked.
    I will not throw my r2 out :) it was just a joke.

    cheers :D
     
    flingin, Jan 24, 2012
    #11
  12. GeoCake

    GeoCake http://ted.ph

    Reputations:
    1,491
    Messages:
    1,232
    Likes Received:
    2
    Trophy Points:
    56
    See that 6 under ~4700? That's all my benches xD

    8.800.6.1000 is the driver for crossfire 6970M also, so that looks like a bench from 2 cards in crossfire, and the result must have glitched upon submission OR was not detected properly. The latter is exactly what happened imo, as futuremark is famous for that.

    Even if he switched one card off that's just not possible
     
    GeoCake, Jan 24, 2012
    #12
  13. long2905

    long2905 Notebook Virtuoso

    Reputations:
    2,443
    Messages:
    2,314
    Likes Received:
    114
    Trophy Points:
    81
    some of them are from this guy too :p
     
    long2905, Jan 24, 2012
    #13
  14. aldam

    aldam Notebook Evangelist

    Reputations:
    39
    Messages:
    446
    Likes Received:
    1
    Trophy Points:
    30
    2720qm + 6970m stock = 3.3k
    2720qm + 6970m 880/1150 = 3.9k

    your score is ok
     
    aldam, Jan 24, 2012
    #14
  15. Dellienware

    Dellienware Workstations & Ultrabooks

    Reputations:
    215
    Messages:
    1,588
    Likes Received:
    61
    Trophy Points:
    66
    that score looks pretty reasonable to me. Have you tried to use any software like gpuz to see if there is any throttling during the benchmarking? If no throttling with furmark, im sure the benchies will run alright though.
     
    Dellienware, Jan 24, 2012
    #15
  16. flingin

    flingin M17x R2 Mafia

    Reputations:
    444
    Messages:
    1,173
    Likes Received:
    306
    Trophy Points:
    101
    no throttling as my card never goes >75'C in crysis 2. So my score is ok :)

    thanks guyz ;)
     
    flingin, Jan 25, 2012
    #16