Ups and Downs in Benchmarking Far Cry 2

by Derek Wilson on 10/30/2008 12:00 AM EST
Comments Locked

16 Comments

Back to Article

  • aguilpa1 - Saturday, November 1, 2008 - link

    A couple of very old 8800GTX's Q6600 @ 3.2Ghz are faster than a single 280GTX on this engine. It is a well optimized graphics engine. I did the last run at 1280 x 1024 for apples to apples since my native resolution is 1920x1080 as opposed to the comparison resolution of 1920x1200.

    Resolution 1920x1080:
    High Quality NoAA 75.01
    Very High Quality NoAA 65.26
    Ultra High Quality NoAA 58.17
    High Quality 4AA 65.78
    Very High Quality 4AA 52.33
    Ultra High Quality 4AA 43.97

    1280x1024:
    Ultra High Quality 4AA 54.64



  • Rock Hydra - Thursday, November 6, 2008 - link

    I don't know why exactly that is, but I remember the thread in the forums showing 2 old ass video cards running Doom 3 in SLI. 3DFx maybe, well anyway it ran at playable frame rates, but quality was bad. So I mean, if your card can't render everything, it will perform faster. Not necessarily the case, but also a thought about why performance might be better.
  • aguilpa1 - Saturday, November 22, 2008 - link

    WTF are you talking about? The quality is the same, and it renders everything same as the 280GTX and the 4870's.
  • marlinman - Friday, October 31, 2008 - link

    The tool IS cool. Not convinced that 3 loops is sufficient though given that during each loop a slightly different scene is displayed. I've noticed (with a little suspicion) that the first loop invariably boasts a higher fps than subsequent loops (in earlier games the reverse invariably holds!). Stability of tool is fine here - with fixed settings, I ran 30 loops (10 for each demo map size) without incident. The first ten results (which are quite representative)

    43.47
    36.67
    35.33
    37.56
    38.67
    36.25
    37.51
    36.59
    38.97
    37.12

    suggest I think that averaging the first 3 runs could be misleading(?)
  • Atechie - Friday, October 31, 2008 - link

    I still don't get the love for the canned benchmark tool...how does the result compare to actually playing the game...or wasn't that a factor?
  • BLHealthy4life - Friday, October 31, 2008 - link

    I cannot get DX10 and AA to work with this game at all! I've tried 178.24, 180.42 and 180.43, NONE of those drivers will give me DX10 and AA at 2560x1600!

    Am I the only one having this f'ing problem?!
  • doomedtofrag - Friday, October 31, 2008 - link

    Hello Derek,

    I understand performance analysis is the primary concern, but it would be nice if you could include in-game screenshots showing quality differences under various settings. Your site does similar thing when it comes to digital cameras (of course that is their primary function). As a gamer, I would love some in-game screenshots, as my faith in the plethora of gaming sites is dwindling day-by-day...
  • ThomasS31 - Friday, October 31, 2008 - link

    Derek,

    Do you plan to investigate the new Stalker game AMD DX10.1 vs nVidia DX10 quality performance situation as well?
    Might worth a short, but good quality article on that. :)

    Tamas
  • ThomasS31 - Friday, October 31, 2008 - link

    Hi Derek,

    Please note (also in the article) that the GeForce DX10 and Radeon DX10.1 card are not doing the same amount and type of calculation with AA enabled!
    So might worth noting that one-to-one compare is not possible or "fair", both on performance and quality.
    (Similar, but not he same like the Assassin Creed case...)

    Regards,
    Tamas
  • thebeastie - Friday, October 31, 2008 - link

    I been looking at getting a new PC for the last 6months at least but just aren't happy with whats out there.
    The main idea behind the new PC is to play games.

    I been waiting for the new Nehalem CPU but I have been thinking I am probably better off making sure I get the latest GPU compared to the new Nehalem.

    Would be good to get a comparison of whats more worth while throwing money at. I think a lot of people suffer from this problem.

  • marsbound2024 - Thursday, October 30, 2008 - link

    Graphically, does it compare with CoD4 or does it fall a bit short of that? Also, going to bed at perhaps 9:00-9:30PM EST? I am guessing you are living in the Eastern Time Zone, though I am just assuming. If it is central that'd make it 8-8:30. Seems a bit early for night owls such as myself. :)
  • DerekWilson - Thursday, October 30, 2008 - link

    it's ... sort of different than cod4 ... i'd say that the visuals are on par, but good for different reasons. if that makes sense.

    i posted this at 12:00:01 am EDT btw, not 9 :-)

    it might not have made it to the front page til later, but it was up in our blog section before then.
  • marsbound2024 - Thursday, October 30, 2008 - link

    Heh, I just assumed that it was an instantaneous appearance on the website when you post an article. To me, the article just appeared like, almost an hour ago (but then of course it just appeared on the front page, as you've said). I suppose there is some sort of delay (perhaps for the editor to take a gander?).

    Anyways, I think I can understand what you're referring to when you say the visuals are on par, but they aren't the same. I am sure the games use a different engine and so the way they are rendered is different. Besides, the atmosphere of the game is expectedly different than CoD4.

    Look forward to more blogs--and hey maybe game devs will take note and start including better bench utilities in their games.
  • Regs - Friday, October 31, 2008 - link

    Particle and volumetric visuals on COD4 are more intense, and the shaders are a little more detailed especially for characters on COD4. FC2 looks a little more waxy on the surface compared to COD4 because of less use of shaders. My guess is because the game is based off the intended use of console hardware.
  • Boushh - Saturday, January 2, 2021 - link

    Always fun to run an old Benchmark on modern hardware:

    Intel i5 9600KF
    16 GB DDR4 at 3200 Mhz
    GeForce RTX 27070 Super
    Windows 10

    Results (2560x1440, DX10 Ultra High settings):

    Average: 257 fps
    Maximum: 387 fps
    Minimum: 177 fps

    The GTX 280 in the article does 30 fps on 2560x1600. If that is the maximum fps, then the modern setup is almost 13 times as fast !!!
  • Boushh - Saturday, January 2, 2021 - link

    That should be '2070 Super' and not '27070 Super' !

    I also ran the same Benchmark on an older Windows XP system (Core 2 Duo E7600, 4 GB DDR2 at 800 Mhz and a GeForce GTX 760), same settings (2560x1440 and Ultra) but in DX9 and it maxed out at about 63 fps. But dipped as low as 25 fps.

Log in

Don't have an account? Sign up now