5870 - DirectX11 vs 10 in battleforge

[straxusii]

http://www.fudzilla.com/content/view/15721/1/

Interesting, no extra eye candy, just fps

 

[straxusii]

http://www.driverheaven.net/reviews.php?reviewid=842&pageid=5

Reckon they got the min framerates on the 295 the wrong way around.

 

[straxusii]

Not really, 2560x1600 4x AA could be pushing the Vram on the ATI card & the 5xxx & DX11 drivers have some maturing to do.

No I mean it shows the 295 as 16 min at 1920x1200 and 21 min at 2560x1600, reckon they are the other way around.

 

[straxusii]

Who cares, it doesn't matter

ROFL!

 

[straxusii]

The thing about minimum framerates though is that they can be affected by transient spikes. Like, say your computer hiccups randomly (accessing a file, computing something in the background etc), it can show up in the benchmark. Such things would be smoothed out in the average FPS, but not neccesarily in the minimum value.

The ideal way to measure the minimum framerate is to run the test several times, and take the "maximum minimum", i.e. the largest of the recorded minimum values. If they just run the test once, or they take the absolute minimum of all the runs, it's entirely possible for a faster benchmark to have a lower minimum framerate for the reasons stated above.

Sure but they are also caused by the graphics card being hammered

What we need in addition to minimum is minimum 95th percentile so sort the results in order and report the result at 5% in.

 

[straxusii]

I have nto tired this game yet I was going to install tonight to see if there was any use of dx11 but are you saying that there is no visual difference between dx10 & dx11?

In this game only there is no visual difference, purely because all the dev has done is to use some DX11 features to boost performance.

Not saying there is no visual dif between DX10 & 11 in general!