The 990FX Chipset Arrives: AMD And SLI Rise Again

Benchmark Results: Metro 2033 (DX11)

Metro is an absolute beast when it comes to taxing graphics performance. Applying the depth of field effect, especially, has a profound impact on frame rates, as each card is hit with a DirectCompute task that literally cuts results in half.

It’s really not a surprise, then, that it doesn’t matter if you’re running a Core i5 at 3.1 GHz or a Phenom II at 3.7 GHz—both solutions yield comparable performance when they’re backed by a pair of GeForce GTX 570s in SLI.

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.
TOPICS
  • Marco925
    But! Does it play metro 2033?


    It Does!!!!
    Reply
  • nice to see support for both videocard producers. especialy for nvidia. now you can do amd+nvidia not only amd+ati(amd)
    Reply
  • nforce4max
    So fast that it sucks your eyeballs into the back of your skull C:
    Reply
  • stingstang
    Tom's, what the hell is this? "At the end of the day, it's the graphics cards which are the bottleneck."
    Did you go about benchmarking graphics cards, or was this a motherboard/cpu comparison? I'm tired of hearing this excuse all the time. We know you have a pair of 6990s and 590s in your shop. Get rid of that stupid bottleneck and DO IT RIGHT!
    Reply
  • wishmaster12
    Of couse you guys know their using Nvidia cards.
    Reply
  • saint19
    I'd keep in mind that this performance review was made it with an AM3 CPU and not with the new generation.
    Reply
  • lavitz1125
    Eh, too late too little.
    Reply
  • geekapproved
    What performance review? They didn't get to test anything. LOL
    Reply
  • -Fran-
    Thanks for the review, but at lower resolutions we all know that the CPU differences will become clear. So you just proved that if a game is taxing on the GPUs, both solutions are equal and when the graphics card ain't being taxed, CPU differences become apparent... Ok, thanks for proving what we already know once more (not being sarcastic here >_
    Reply
  • -Fran-
    Uhm... that last comment of mine was cut in half with the missing char of that face... I guess that's an escape char; oh well.

    What is missing said something like:

    ...here "face"), but you said you wanted to test AMD's SLI on their 990FX vs Intel's SLI. So, IMO, you need less graphics horse power: like 2 GTS250's or 2 GTX460's or 2 GTX560's (not ti's) to tax the graphics subsystem and really show the differences. Maybe up the resolution also to really show if there is a difference between AMD's or Intel's SLI.

    Thanks again for the Article, Mr Chris.

    Cheers!
    Reply