I did some benchmarking on XP x64.
2 instances, windowed @ 1280*800:
Frames: 2488 - Time: 30000ms - Avg: 82.933 - Min: 26 - Max: 185
1 instance, windowed @ 1280*800:
Frames: 4201 - Time: 30000ms - Avg: 140.033 - Min: 58 - Max: 273
1 instance, fullscreen @ 2560x1600:
Frames: 4432 - Time: 30000ms - Avg: 147.733 - Min: 61 - Max: 308
Tests are somewhat similar regarding Nvidia settings - LFS settings were the same.
Not too bad as the results are pretty close to each other across OS'. Again.. 32 bit vs. 64, meaning I'm not expecting 100% match.
I also did some
Unofficial LFS benchmarks, and my ranking seems to be OK, however I do see some lower speced systems ranked above mine (CasseBent.. what are you doing up there?
or how, really)
Overall LFS performance never really bothered me. It's this multiple instances thingy, so I tried to do a bit more controlled tests:
An attemt at compareable benchmarking
XP 32 bit, default Nvidia settings. LFS settings set to max, incl. AA and AF
2 instances, windowed @ 640x480
Frames: 2310 - Time: 30000ms - Avg: 77.000 - Min: 34 - Max: 154
1 instance, windowed @ 640x480
Frames: 4524 - Time: 30000ms - Avg: 150.800 - Min: 70 - Max: 298
1 instance, fullscreen @ 640x480:
Frames: 4588 - Time: 30000ms - Avg: 152.933 - Min: 74 - Max: 296
I didn't get to do the tests under similar circumstances on XP 64. I unlocked on XP 64, then unlocked same LFS on XP 32, and apparently that ruined my unlock on XP 64.. meh. Doesn't matter too much - the reason i did this was because I would like some of you to give it a try with similar settings - that way it's much much more comparable:
1. reset gfx board settings to default.
2. use the attached LFS config file (remember to rename/backup your own first!)
3. Configure Fraps to do a 30 sec benchmark (you can
get Fraps here).
4. run 2 instances of LFS and place them side by side - don't alter any setting/option.
5. use the attached replay.
6. change to the very last car (shift+tab once). Don't change cam angle.
7. let the replay run once on each instance. I don't know if it helps to avoid hick ups, but won't hurt.
8. rewind to the beginning on both instances, and then start them immediately after each other.
9. start the benchmarking when the lights goes green - don't do anything the next 30 secs.
Fraps will stop after 30 seconds - you can check the result in the FRAPSLOG.TXT file, which is located in your Fraps folder.
10. while both LFS instances are still running (but after the benchmarking is done) - try using other apps such as a browser, to see if there are any slowdowns.
11. make sure all LFS instances are closed before you restore your original config file!
12. post your results/findings here, including your basic computer specs such as CPU, frequency, gfx board and OS.
Would be interesting to see the results, but I'm suspecting that my LFS in multi mode are running fine. The replay is much heavier than what I normally see/am used to so I'm probably very biased.
Video guide of how to perform the test.
Few notes about the video:
The video was recorded at 30 FPS, but is lagging. Fraps can do a much better job even at 2560 res, but Fraps can't record non-OpenGL/DirectX stuff.
At the end of the video/benchmark procedure, you see me using 2 apps. Although it may seem that the lagging is a result of the recording, the lag is very much like how it really is.
I then closed one instance and "used" the apps again. This time its much faster - you still see a bit update lag, but this is entirely caused by the recording software. The lag/delay is virtually none under normal circumstances.
There are no slowdown on XP 64. Here I can run 12 instances of LFS, all placed side by side, and although there is a bit of slowdown when running that many, it's still far from what I'm experiencing on XP 32 with just 2 - on XP 32 it can even be worse than what you see in the vid.
Ps. it's fine if you only do the benchmark using 2 instances - thats the one that really matters anyways.