Great review, although you may want to upgrade your test platform to prevent possible CPU bottlenecking with Tri-SLI GTX570s. HardOCP found significant gains by swapping a Core i7 920 @ 3.6ghz for a 2600k @ 4.8ghz: http://www.hardocp.com/ar...y_sli_amd_trifire_redux/2

Also, it makes sense that someone who is spending $1000 on GPU hardware is going to pair it with the fastest possible CPU platform (so it's hardly realistic that this gamer will use a stock i7 975 @ 3.33ghz). Just saying

I wish you would keep the Crysis / Warhead tests. We have years of experience and comparative data for those, and they can still do a great job in testing PC gfx capabilities. Much better than Crysis 2, where you already have to go to 2560x1600 to see a significant difference in frame rates. Finally, IMHO Crysis / Warhead are better FPS games than anything that has come along since. They deserve to stay on the charts.

Its good to see varied and extensive lineup of games in your reviews. Games like Shogun 2 Total War are not benchmarked commonly by other sites.

One thing you may like to check and be wary of with regard to Shogun 2 is the 'Downgrading algorithm' which automatically downgrades the graphics of the game if it feels your graphics card lacks memory. Happens to cards with 1280MB or even 1536Mb. This may result in skewed and inaccurate results in favour of cards with less memory.

You can check the log file in the 'Creative Assembly' folder for any such downgrading. To remedy this, you can put in an override in the 'preference.txt' file in the scripts sub-folders by replacing 'gfx_video_memory 0' with 'gfx_video_memory 2147483648' (for 2GB or twice as much for 4 GB); save and exit.