Radeon HD 7990 Vs. GeForce GTX 690: The Crowd Picks A Winner

Results: Grid 2

Without frame pacing enabled and running at 5760x1080, AMD clearly has some issues in Grid 2. This was evident in Don’s data at 2560x1440, too, where the Radeon HD 7990 suffered a number of performance dips down into territory that was certainly palpable. We’re able to use the word certainly because nobody preferred the 7990, two gamers couldn’t tell a difference, and four unknowingly chose the GeForce GTX 690.

Frame pacing appears to add notable consistency to the 7990’s performance. We see it in the subjective results, as most of our participants didn’t prefer one card over the other, one picked AMD, and two chose Nvidia.

We can also switch over to Don’s numbers at 2560x1440 and see that, although frame pacing knocks some of the peaks off of AMD’s previous drivers, by also eliminating the extreme performance dips, average frame rates persist, as the minimums come way up. Frame time variance shrinks, which is probably why more of our attendees leaned closer to AMD after the first part of our experiment.

Chris Angelini
Chris Angelini is an Editor Emeritus at Tom's Hardware US. He edits hardware reviews and covers high-profile CPU and GPU launches.
  • iam2thecrowe
    can you please show cpu useage before and after the driver update, and also show results on a lower end system, like an fx6300 with dual 7850's? That would be interesting.
    Reply
  • bin1127
    A great article with some very interesting results.

    Chris really knows use funny and imaginative metaphors to start of his articles. Keep up the good work!
    Reply
  • iam2thecrowe
    can you please show cpu useage before and after the driver update, and also show results on a lower end system, like an fx6300 with dual 7850's? That would be interesting.
    Reply
  • merikafyeah
    If I had $700 to spend on graphics, I'd buy the fastest single-gpu card there is:
    The Galaxy GeForce GTX 780 HOF Edition

    http://www.hardocp.com/article/2013/08/14/galaxy_geforce_gtx_780_hof_edition_review/
    http://www.amazon.com/dp/B00DZIFN4M/

    Out of the box it beats the Titan in pretty much everything and when overclocked I'd bet performance would be indistinguishable from a GTX 690. And did I mention single GPU? Frame-pacing = non-issue. And those temps/noise ratios are in a league of their own.
    Reply
  • iam2thecrowe
    11482712 said:
    If I had $700 to spend on graphics, I'd buy the fastest single-gpu card there is:
    The Galaxy GeForce GTX 780 HOF Edition

    http://www.hardocp.com/article/2013/08/14/galaxy_geforce_gtx_780_hof_edition_review/
    http://www.amazon.com/dp/B00DZIFN4M/

    Out of the box it beats the Titan in pretty much everything and when overclocked I'd bet performance would be indistinguishable from a GTX 690. And did I mention single GPU? Frame-pacing = non-issue. And those temps/noise ratios are in a league of their own.

    you are pretty much right on there. people often overlook galaxy cards (i know i do), for obvious reasons (quality/after sales support) . but this card is insane, with extremely good OC capabilities. If i had a choice between this and a 690, i would certainly take this card. Althogh i would take 2 x of the cheapest 770's available before any of the mentioned cards.
    Reply
  • spentshells
    How was the beer?
    Reply
  • AndrewJacksonZA
    I think the most important part of this article is this little gem:
    "My wife is in the background with our then-two-week-old baby girl"
    Congratulations Chris!!! :-)
    Reply
  • geok1ng
    With due respect for TH efforts on the subject, but keeping Metro LL results is a BIAS. How about some piecharts WITHOUT METRO LL?
    Reply
  • Bill Reinhardt
    You know how wine is matched for optimal flavor compatability with certain foods?

    http://en.wikipedia.org/wiki/Wine_and_food_matching

    You should have done that with beer and video games.
    Reply
  • Iron_Sean
    Why did you only test Triple monitors with no frame pacing versus single high res monitor with frame pacing?

    You're confounding variables and making your results hard to compare directly: Is the difference 6 MP vs 4 MP, frame pacing, or their crossfire/surround technology? This is an interesting idea, but I'd really like to see it narrowed to a single variable to better compare the technologies.
    Reply