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

In Pictures: Tom's Hardware's Graphics Testing Event

The first attendees arrive as we finish setting up the room...

Here I am, answering questions and explaining the purpose of what we're trying to do at the event.

A local redditor sits down to Grid 2 for A-B testing.

More blind game testing on our black and white test platforms. My wife is in the background with our then-two-week-old baby girl (the reason our Catalyst 13.8 coverage took a little longer to put together).

More analysis of Grid 2, trying to decide which experience is smoother.

Our exhibition systems. From left to right, CyberPower's Zeus Hercules with Iris Pro, iBuyPower's FX-9590-powered setup, and Tom's Hardware's Core i7-3970X-based water-cooled system with dual GeForce GTX Titans.

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