Benchmark inconsistencies?

[Amblin]

Hello,

Hi, I've been lurking and downloading from here since Rome 1, but have now registered. Greetings and thanks to the community here for so many ace mods throughout the good and bad times!

Now that is out of the way, can I ask, is anyone else experiencing inconsistency in the "Forest" Benchmark?

I have ran the benchmark a number of times this morning and have noticed that the benchmark isn't actually running the identical test each time.

It is my finding, proved again after running it over and over again that when the "Hay bales of fire" roll down the hill toward the 2 Roman units, and the camera pans right, the same outcome, is not consistent. I have found, without changing settings that...

1) Sometimes soldiers are set alight, sometimes not. There may be variances of 1, 3 or 5 set alight.
2) Sometimes only a single hay bale rests on the rocks, sometimes 2.
3) Sometimes a single hay bale rests in the "nook" of the rocks and sometimes it rolls away.

There may be others, but my point is, if the same scene isn't used for benchmarking, benchmarking is unreliable. I suspect there is a random element to the hay bales trajectory that's causing this.

It's not my imagination, play around, change some settings, you may find the same!

If it's of interest, I'm on an i5 2500k at 4.6 with 16GB RAM and AMD/ATI 7970 @ 41-43 FPS in the benchmark, depending on what happens! (Shadows on High all else on Extreme)

Cheers and good luck all!

Amblin

[The_Dude]

you do realize this isnt in any way, shape or form a "standardized" benchmark ?
this isnt 3dmark or anything, its just there for you get a rough idea of which settings you can run ... if that is a couple of fps off in either direction cause theres minimal(and what you describe is minimal) random elements it really doesnt matter.

[Amblin]

A valid point Dude, but where small changes to a single setting may net a 5fps+ gain, I felt this was important enough to ask about.

Cheers!