This week I took my XP-based gaming rig and decided to install Vista on it. Even though I'm running Vista on my other machines, I decided not to touch this machine because it was mainly used for gaming, and early reports were that the framerates in Vista were pretty bad compared to equivalent XP machines.
I decided to dual boot so I could test this myself. Up first: Half Life 2/CS:S. I ran the default benchmark -- I set the resolution on 1680x1050, no anti-aliasing, all other options on high. My CPU is a dual core AMD Opteron 180, with an old NVidia AGP 6800GT. I'm using the latest drivers as of this date.
When I first ran the test, the XP machine was benching much better, but then I realized I had RivaTuner overclocking the video card on XP. With all that normalized, the XP machine was able to outperform the Vista machine, but only by 2 fps. (81 fps vs 79 fps.) This is hardly a noticeable difference and certainly within a margin of error.
Next I tried Company of Heroes. This is a very demanding RTS, and one that was getting horrible performance scores in Vista when it was first released. I kept the resolution high as well (1680x1050) but most other settings were set to low. Here, Vista actually outperforms XP, yielding an average of 50.4 fps on Vista compared to 48.6 on XP. Again, only a 2 fps difference.
Obviously, these tests aren't completely comprehensive. But it's a good starting point, and proved to me that Vista is just as capable of gaming as XP; in fact, I think over time as dual core chips become all the rage and multithreaded application design and 64-bit architecture becomes the norm, Vista will pull ahead.
So why the bad numbers early on? Most likely the driver optimizations weren't there. As migration has continued and hardware vendors have had time to tune their drivers for the new Vista driver model, the performance has improved.
Have another test you want run? Let me know, I'll see if I can do it.
Subscribe