I'm still waiting for my copy of Battlefield 3 to finish installing and updating and all that, but as soon as it does I'll be putting to the test the many claims of hardware vendors that you need to spend a fortune on new kit to get it playing perfectly.
While I'm waiting, however, here's a little experiment I'd like you all to try at home.
There's no way I can benchmark every combination of kit available to see how it performs running Battlefield 3 in the lab. But you want to know which graphics processor you should buy if you own x , or how much better y CPU is going to be than z .
So here's a better idea: discuss amongst yourselves. With our help, of course.
We want as many of you to benchmark your PC or laptop running Battlefield 3 as possible, so that we can create the most comprehensive test of a program's framerates ever. If it works, you should be able to compare your performance with similar set ups, and figure out the best upgrade by finding other combinations of kit already tested.
If it doesn't I'll delete this post and we shall never speak of it again.
It's entirely unscientific, since proper testing would require everyone to be running the same version of Windows and the same drivers and using exactly the same sequence to benchmark in. But as a rough guide to what's good for performance, it's more than just a bit of fun.
All you have to do to take part is install Battlefield 3 and FRAPS . Start FRAPS first, and set it to record the minimum, average and maximum framerates (Min/Max/Avg), bound to your favourite hotkey. Then, when you're in the game press the hotkey once to start the test, and press it again a minute or two later to stop it. Check the log in your FRAPS folder (usually C:\FRAPS), and enter the results, along with some hardware information, in the form below.
To try and keep things roughly comparable, if you could perform the test sometime around the start of Operation Metro, where you're running around and fighting outside, that keeps things simple and more or less the same. Needless to say if you die while benchmarking you'll probably need to start again after you spawn.
The important thing is that we're all using the same settings and - as far as possible - resolution. So if you could use a generic 'High' profile for graphics and submit one set of scores at 1920x1080, that'd be much appreciated too.
If you want to share the results on your triple-monitor 3D capable machine for bragging rights, that's fine - but getting as many as possible at the same resolution is what will make this work. So feel free to run the test twice. If you have a smaller screen, then you can also take part but just make it clear in the question about res.
Once results start to come in, I'll publish a live version of the spreadsheet as a separate post. If you're really keen, though, you can see the rawdata by following this link .
I should point out that no personal details are collected in this Q&A or held by myself or Future Publishing. It's all as anonymous as anything on the web can be.
So get benchmarking and fill in the form below. And when you've done so, invite your friends to do the same.
