ZOTAC NVIDIA GTX Titan Graphics Card Review @ HH

Discussion in 'Reviews & Articles Discussion' started by craig5320, Feb 26, 2013.

  1. craig5320

    craig5320 Well-Known Member

    Joined:
    May 7, 2002
    Messages:
    9,682
    Likes Received:
    226
    Trophy Points:
    88
    ZOTAC NVIDIA GTX Titan Graphics Card Review

    "Today we take the GTX Titan through a selection of real world gaming tests which include Crysis 3, Aliens: Colonial Marines, Far Cry 3, Assassins Creed 3, Hitman: Absolution, StarCraft 2: Heart of the Swarm, Battlefield 3, Skyrim: Dragonborn, DOTA 2, SWTOR, Mass Effect 3, F1 2012 and Borderlands 2. Mix that with the latest 3DMark, Heaven Benchmark, GPU computing, Media playback and all the usual power/thermal and overclocking tests... add frame latency, competing multi and single GPU setups and we have quite a lot to get through today."
     
  2. Erroneus

    Erroneus Get off my lawn!

    Joined:
    Aug 25, 2002
    Messages:
    13,727
    Likes Received:
    101
    Trophy Points:
    103
    I agree with the conclusion, impressive card, but the pricing is off completely.
     
  3. IvanV

    IvanV HH Assassin Guild Member

    Joined:
    Dec 18, 2004
    Messages:
    10,840
    Likes Received:
    2,063
    Trophy Points:
    138
    Cool, I'm glad that you're testing it and I also like the addition to the testing methodology. :)
    I don't think that this is meant to sell, this was made to grab a few headlines and claim technological superiority and the price is deliberately set too high for the card be able to generate any significant demand. That way, NV doesn't have to make too many to be able to claim availability (and therefore that this is a real product) and the Titan won't mess with the sales of the actual product lineup. Still, a very cool piece of kit.

    I'm still not sold on FXAA or MXAA. They are good technologies, but have been "caught" decreasing texture clarity. Being an AMD user, I prefer to enable something like 2xEQ AA and adaptive multisampling, for really nice and fairly low impact smoothing.

    BTW, on page four, the bars in the Heaven graph are all of strangely similar length. For example, the single 7970 scores almost exactly half of what 7970 CF does, but the bars are almost the same.
     
  4. malmental

    malmental System Builder > Hardware Specialist

    Joined:
    Sep 17, 2012
    Messages:
    1,589
    Likes Received:
    49
    Trophy Points:
    48
    more FPS on 690 isn't news, but using 0x AA and 4x AF shows exactly why the 690 beat everything else so handily.
    VRAM has great use for triple screen. Single screen? Not so much

    try using some real extreme settings for benching.. :rolleyes:
     
  5. IvanV

    IvanV HH Assassin Guild Member

    Joined:
    Dec 18, 2004
    Messages:
    10,840
    Likes Received:
    2,063
    Trophy Points:
    138
    The single screen tests used maximum available setting. In the situations in which AA was disabled for the three-screen test, the Titan was averaging thirty-odd frames per second and dipping into twenties. Would it outperform the 690 if the settings were more extreme? I don't know, but it sure wouldn't be playable!
     
  6. craig5320

    craig5320 Well-Known Member

    Joined:
    May 7, 2002
    Messages:
    9,682
    Likes Received:
    226
    Trophy Points:
    88
    The scale is logarithmic in order to have the fps register on the chart, I really should have added the bottom scale so that was apparent, I've now updated the chart.

    Thanks
     
  7. IvanV

    IvanV HH Assassin Guild Member

    Joined:
    Dec 18, 2004
    Messages:
    10,840
    Likes Received:
    2,063
    Trophy Points:
    138
    Aah, I get it... I thought "Why on Earth would you use logarithmic scale?!", but than I realized that the fps and the points values had to share it. Thanks for the clarification. :)
     
  8. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    Yup, spot on. We were aiming for pretty close to maximum playable settings.
     
  9. Takaharu

    Takaharu Unus offa, unus iuguolo

    Joined:
    Aug 31, 2009
    Messages:
    3,368
    Likes Received:
    374
    Trophy Points:
    108
    Shouldn't you have used three-four 7970s for comparison to match the price range? :lol:

    Nice work on the review, lads. Interesting spikes on the frame times. Can anyone explain the vast amounts of oscillations?
     
  10. blibbax

    blibbax nahm8

    Joined:
    Sep 6, 2008
    Messages:
    6,210
    Likes Received:
    203
    Trophy Points:
    73
    Great job on the new performance metric, which to me is far more useful than FPS numbers.

    If you think about it, what's happening on-screen in any one frame is very variable even during the course of a second. Even with perfect drivers, you'd expect some variation. Thing is though, drivers are far from perfect, and I think that's most of what causes this.

    Certainly Titan is not an answer to the latter. Like any luxury product, Titan is designed to do (at least these) two things:

    • Establish Nvidia's reputation as king of its game
    • Take big wadges of cash from money-no-object buyers
    Think about it like a Rolls Royce. It isn't as good as 20 Volkswagen Golfs, and a large number of buyers probably own half a dozen other similarly priced cars already.

    In addition, Titan serves as a relatively low cost alternative to Tesla for those that don't need ECC.

    I don't think Nvidia could satisfy demand for GK110 chips if they made the price any more accessible (yet).

    I actually feel like most Titans will end up being used in SLI. This is not a product for the price conscious, and if you "only" have £1000 to spend, a 690 will often make more sense.
     
  11. beck2448

    beck2448 New Member

    Joined:
    Mar 1, 2013
    Messages:
    2
    Likes Received:
    0
    Trophy Points:
    0
    There is a very interesting article today at Tech Report and PC perspective.
    The writer says that Crossfire has a problem with runt frames, frames that are just slivers, not fully rendered, but fraps counts them as if they are visible frames. When these so called runt frames are removed Crossfire is no faster IN REALITY than a single GPU. This could explain why many sites have noted that despite higher fps numbers Crossfire wasn't smooth. if this proves to be true and widespread it is a major story and should force the industry to become more exacting in how they measure GPU performance. Its also important that reviewers not take these fps numbers at face value until this matter is investigated.
     
  12. blibbax

    blibbax nahm8

    Joined:
    Sep 6, 2008
    Messages:
    6,210
    Likes Received:
    203
    Trophy Points:
    73
    I also saw that article - http://techreport.com/blog/24415/as-the-second-turns-frame-captures-crossfire-and-more

    If it's true that all Crossfire games are, as you put it, "runt frames", then the hardware reviewing community certainly needs to take a long, hard look at itself. I suspect that the true outcome isn't actually that extreme, but I think that Techreport has shown recently that graphics card reviewers have been somewhat complacent in recent times, missing important results in the process. If Techreport think they can improve their process yet further by analysing what goes down the DVI cable itself, I'm all in favour.
     
  13. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    This is something I discussed with "a major player in the GPU industry" just 2 days ago... how we look at what comes out of the cards and how important it now is.

    Stay tuned ;)
     

Share This Page

visited