Civilization V

The other new game in our benchmark suite is Civilization 5, the latest incarnation in Firaxis Games’ series of turn-based strategy games. Civ 5 gives us an interesting look at things that not even RTSes can match, with a much weaker focus on shading in the game world, and a much greater focus on creating the geometry needed to bring such a world to life. In doing so it uses a slew of DirectX 11 technologies, including tessellation for said geometry and compute shaders for on-the-fly texture decompression.

We have always considered Civ 5 an interesting game both for its near-complete use of the DX11 feature set, and because of its interesting performance characteristics. 2 weeks ago we called it CPU limited based on the fact that once we had sufficiently powerful cards, AMD and NVIDIA results tended to clump together despite any difference in their respective cards’ speed. With the Forceware Release 265 drivers, NVIDIA has blown this assumption apart, with NVIDIA’s more powerful cards launching ahead at 1920 and lower. We appear to be quite GPU limited on the NVIDIA side all of a sudden, which is about as drastic change as we could expect. Furthermore NVIDIA is holding their cards close to their chest on this – they’ve obviously found a wonder optimization, but they aren’t ready to say what it is.

In any case while AMD has always trailed NVIDIA in single card performance in Civ 5, with these driver changes it’s quite monumental. The GTX 560 Ti is 44% faster than the 6950 at 1920, 80% faster at 1680, and even the GTX 460 768MB can edge out the 6970 at 1920. Whatever NVIDIA has done, it has made Civilization V quite a lot faster and smoother at 1920 and 1680, particularly when a large number of units are on screen.

Among NVIDIA’s own cards the lead has actually shrunk some compared to our earlier games. The GTX 470 has an edge on the GTX 560, and the 560 in turn is down to a 25%-30% lead over the GTX 460 1GB. We don’t seem to be ROP or memory bandwidth limited, so perhaps this is a case of the GF104/GF114 architecture’s shaders underperforming?

HAWX Battlefield: Bad Company 2
Comments Locked

87 Comments

View All Comments

  • Nimiz99 - Tuesday, January 25, 2011 - link

    One of my buddies has a C2D 8500 system OC'd to 3.5 i think. He got himself a 5870 (overclocked) to game. The problem we ran into was that the C2D is too slow to handle games like Civ5 that heavily rely on the CPU to keep up (you can still play the game, but it's literally wasting the 5870 with noticeable lag from the chip). Basically, he is upgrading now to a sandy bridge. I'd wager some of the older i7's or maybe even a Thuban (OC'd to 3.8 with a good HT overclock) could manage, but why bother when a new architecture is out form Intel (or AMD later in the year).
    So enjoy your new build ;),
    Nimiz
  • Beenthere - Tuesday, January 25, 2011 - link

    Over the last couple years Nvidia has really struggled and they may be on the ropes at this point. They have created a lot of their own problems with their arrogance so we'll see how it all plays out.
  • kilkennycat - Tuesday, January 25, 2011 - link

    eVGA GTX560 Ti "Superclocked" Core: 900MHz, Shader 1800MHZ; Memory 4212MHz $279.99

    ~ 10% factory-overclock for $20 extra, together with a lifetime warranty (if you register within 30 days) ain't too shabby....
  • Belard - Tuesday, January 25, 2011 - link

    Sure, the name shouldn't be a big deal... but each year or worse, Nvidia comes up with a new marketing product name that is meaningless and confusing.

    Here is the full product name:

    GeForce GTX 560 Ti But in reality, the only part that is needed or makes ANY sense is:
    GeForce 560

    GTX / GT / GTs are worthless. Unless there were GTX 560, GTS 560 and GT 560. Much like the older 8800 series.

    TI is only added to this idiotic mess. Might as well Ultra, Pro or MX.... so perhaps Nvidia will come out with the "GT 520 mx"?

    The product itself is solid, why turn it into something stupid with your marketing department?

    AMD does it right (mostly), the "Radeaon 6870" that's it. DUH.
  • omelet - Tuesday, January 25, 2011 - link

    Yeah. Not that it really matters. And while this might be what you meant by "mostly" note that AMD's naming was pretty retarded this generation with the 68xx having lower performance than 58xx.

    But I don't see why they readopted the Ti moniker.
  • Sufo - Wednesday, January 26, 2011 - link

    no, that's only a result of the 5xxx series being stupidly named. Using 5970 for a dual chip part was the error. Use an x2 suffix or smthng. AMD is back on track with the 6xxx naming convention... well, until we see what they do with the 6 series dual chip card.
  • Belard - Thursday, January 27, 2011 - link

    The model numbers of:

    x600, x800, etc have been consistent since the 3000 series.

    x800 is top
    x700 is high-end mid range ($200 sub)
    x600 is mid-range ($150 sub)
    x400~500 low-end ($50~60)
    x200~300 Desktop or HTPC cards.

    AMD said they changed because they didn't want to confuse people with the 5750/5770 cards with the 6000 series. Which is completely stupid... so instead they confuse everyone with all th cards.

    If the 6800s were called 6700s - they would have been easily faster than any of the 5700s and at least somewhat equal to the 5800s (sometimes slower, others faster). Instead, we have "6850" that is slower than the 5850.

    The prices are a bit high still, yet far cheaper than the 5800 series, in which a 5850 was $300+ or $400 for the 5870. But by all means, I'd rather spend $220 on a 6870 than $370 on todays 5870s.

    Anyways, I'm still using a 4670 in my main computer. When I do my next upgrade, I'll spend about $200 at the most and want at least 6870 level of performance, which is still about 4x faster than what I have now. Noise & heat are very high on my list, my 4670 was $15 extra for the better noise & heat cooling system. Perhaps in 6 months, the AMD 7000 or GeForce 700 series will be out.
  • marraco - Tuesday, January 25, 2011 - link

    Is the first time I see a radiator geometrically aligned to the direction of air velocity thrown by the fan.

    Obviously it increases the efficiency of the fan, increasing the flow of air thrown across the radiator, and reducing noise.

    It’s an obvious enhancement in air cooling, that I don’t understand why CPU coolers don’t use.
  • strikeback03 - Tuesday, January 25, 2011 - link

    I wouldn't be surprised if in some cases the increase in fin surface area (from having a bunch of straight fins packed more closely together) produces better cooling than having a cleaner airpath.
  • MeanBruce - Wednesday, January 26, 2011 - link

    You should check out the four Asus Direct CU II three slot radiators that came out today on the GTX 580, 570, and the HD 6970 and 6950, each using two 100mm fans, five heatpipes and three slots of pure metal, they claim you can easily fit two of them on ATX for SLI and CB?

Log in

Don't have an account? Sign up now