As far as I am aware you cannot increase the max fps limit on the Need For Speed games, and I know you can't in MW3/MW2.
Should def be something representing last 3 generations of DirectX as a lot of games are still coded in DX9 and quite a few readers will still have DX10 cards.
I'd like to see one game that's not so cutting edge in there, since I often pick up a game a year or two after release.
Fwiw Sleeping Dogs actually heats my card up more than furmark or Ungine Heaven, and comes with a built in benchmark.
Arma 2 lol. JK On a serious note, though, bench the original Crysis. Just because Crysis 2 is newer doesn't mean its better. Crysis 1 is still more taxing!
Rule out anything with AMD, Intel NVIDIA splash screens or PhysX You need to use more than three games to spot AMD or NVIDIA bias. Don't be scared of synthetic benchmarks, they tend to scale in the same way as meta scores taken from a broad range of games. I'd also like to see 3D and Multi-Monitor modes routinely benched for top tier cards and multi GPU systems since we've been at a point where any £200 card will handle almost any game at 1080p for a while now and 80-100% scaling is now the norm for Intel based systems.
I would like to have some older titles, such as Team Fortress 2, Borderlands and Dirt2. I know, they're getting silly FPS nowadays, but they do help with comparisons
As much as i'd prefer it, Crysis can only really make use of 1 core so it's not really a fair representation of today's machines, It also has tessellation, even if it is poorly implemented.
Crysis 2 with DX11 and high texture update at 2560x1600 and 16x AA BF3 at same settings. Supreme commander: 8 player with 1000 units and shadows on. Wait never mind that would be a CPU test
For those asking for guild wars 2 it's not gonna happen. Unless there's a section of the game that's on rails ( aka wow flight points) they could never repeat the data otherwise so the comparison card to card would be impossible.
Any source engine game gets 100+ fps on a modern card 560ti onwards passed 60 is pointless. All strategy games are CPU capped.
Shogun 2 isn't, it limits the detail based on what it thinks your card can do, but i think the benchmark plays it at full detail which has been known to bring cards to their knees.