5 fps difference between all medium and high (with medium textures and terrain)

[Unfocused]

I did a benchmark with FRAPS - I joined a Wake server, first with all medium, then with all high (except for terrain and textures on medium). I played for a few minutes, then checked the file that was created and counted the average, maximum and minimum fps that I got with each setting.

1024x768
[avg]
[max]
[min]

On all medium I got:
34,36645963
58
23

On all high except for textures and terrain on medium I got:
29,55590062
49
18

My specs:
PIV2,8Ghz
1,5GB RAM
GF6800GS (128MB/128bit version@508/1300; 4100pts in 3DMark05)
Forceware 84.21

Why is the difference so small? Shouldn't my FPS be higher in both cases anyway? My friend that has the same card, gets about 15fps more (though he's got a score of 4500 in 3DMark05 without OCing it [450/1200], for some reason)

[Jofaba]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

Could be that you started your first run with the lower settings, then increased your settings, then did your second run without clearing out your shaders. That could bog down your peformance a little, and possibly even a lot if you've been changing your settings a lot.

Go here:

C:\Documents and Settings\Administrator\My Documents\Battlefield 2\mods\bf2\cache

Do you see a lot of folders with long code names? delete them all and reload a map. It'll take a while for the shaders to rebuild, then you're playing with fresh files and should get better fps.

Note: The things in bold may be different for your system. C is the drive letter my My Documents folder is in. And "Administrator" is the name of my user account on Windows. If you're not sure, click "Start" and you'll see "log off ... " and a name. That's the name you'll replace Administrator with.

[Unfocused]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

I know the cache trick. I admit that I haven't done that while benchmarking, but I do it almost everytime. I can tell that the fps difference would be small (if any) cause I've been also simply checking FPS while playing on both settings (clearing cache between them) and the fps differences were small, which would match the actual differences shown by benchmarking.

[Jofaba]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

I think you just said the same sentence 3 different ways in a row. If not, then I have no idea what you just said. Sorry.

[Unfocused]

Re: 5 fps difference between all medium and high (with medium textures and terrain)



What I said is: I haven't cleared my cache while doing my benchmarking.

However, I almost always clear my cache, when I change video settings.

The results I get on medium and high settings still vary by those average 5 fps, regardless of whether I clear my cache or not, judging by the in-game fps counter.

Maybe not clearing my cache would affect my performance noticeably if my shaders were optimized many times. While doing the benchmark the shaders were optimized only twice, as I changed my settings once. Thus, it did not cause an fps penalty.

Having said that, I think that the problem lies elsewhere.

[ConscriptVirus]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

there are a lot of settings where between high and medium there sint much diff like geometry and effects. does ur game look that much diff between high and low? and also, the benchmarks were prob not 100% the same so its not that accurate. standing still on a map will get u nice fps and b about the same for all graphics levels

[Unfocused]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

I know that standing still will give me similar fps, that's why I made fraps register my fps and save it to a file and did some running around and battling around the south base, beach and airfield flag, for a few minutes.

By simply looking at the fps (without saving them to a file ie. benchmarking) I can tell there's no big difference between medium and high.

I had an overclocked radeon 9550 before (let's say it was a 9600pro, same chipset, similar clockrates after OC). I played all medium except for lightning and the dynamic stuff (though I sometimes turned them to medium to get more eye candy, but a little less performance). Now that I got a 6800, I expected high settings to give me fps such as I have now, but I thought that medium will be completely comfortable and give me more fps than I had, no slowdowns etc.. I guess I do have more fps, but not much. The game still can slow down noticeably, and I don't feel much difference tbh.

[Unfocused]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

BUMP

Anyone else got any ideas?

[Viper031]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

Nope!

[Hugh G. Wrection]

Re: 5 fps difference between all medium and high (with medium textures and terrain)

To get a meaningful FPS measurement you need to do this:

1) Put 63 bots in AIDefault and then run a single player Karkand game

2) Get yourself down to Hotel where all the action is and start a demo recording. You want to get a good demo with a lot of action going on (so you have a lot of dynamic objects being rendered in addition to static)

3) Repeat until you have a demo that is at least 60 seconds long (ie. you need to stay alive in the hot zone for 60 seconds)

4) Use built-in timedemo functionality and FRAPS. Make sure FRAPS starts benchmarking immediately after pressing the Join button

5) Repeat same timedemo test using different graphics settings (so that your comparisons are valid)