#1
|
|||
|
The LFS Benchmark - Download/Discussion
Hello
Made this little something since LFS has no benchmarking system so far. In short, this is a framerate test. I hope it does the job well enough It should work well on the demo too since the track is blackwood and the cars are GTT's.
Quote:
I hope I covered everything! Thanks to thd, Moonclaw and others in IRC for hints/comments. I'd be neat if you'd post the results here: http://forum.racesimcentral.com/sho...624#post1529624 File: Last edited by vari : 03-02-2004 at 18:12. |
#2
|
||
|
I hope to see many people trying this so we get a good ammount of data for comparison!
Nice work Vari
__________________
The LFS Endurance League | Forum LFS Allround Rank: -0:36.140 (50 %) | +1:29.460 (100 %) GPL Rank: Working on it.... |
#3
|
||
|
i will try, its a great idea. hope everything works fine.
and in some hours i have upgraded my pc....will be interesting, checking the results Last edited by rally : 03-02-2004 at 18:18. |
#4
|
||
|
I think this deserves some attention. Thread stuck
__________________
Michiel Brinkers | RSC Subsites Admin | LFS forum moderator |
#5
|
||
|
Very interesting vari, good work, gonna try soon
|
#6
|
||
|
ups sorry, havent read it right
Last edited by rally : 03-02-2004 at 20:08. |
#7
|
||
|
pssst... don't post results here...
|
#8
|
||
|
Btw, we could come up with a simple rating figure by taking the average of the amount of played frames on min and max settings. In the case of Vari's benchmark that would be: (4945 + 12052) / 2 = 8499 (rounded upwards). Thumbs up anyone?
__________________
Michiel Brinkers | RSC Subsites Admin | LFS forum moderator |
#9
|
||
|
Well some of the point with having 2 alternatives was to show that LFS can be playable even with less than top systems. And a low-end system would come out really bad if both tests were included.
__________________
The LFS Endurance League | Forum LFS Allround Rank: -0:36.140 (50 %) | +1:29.460 (100 %) GPL Rank: Working on it.... |
#10
|
||||
|
Quote:
Good idea except the difference in running time causes some variation because it has to be started/stopped by the user. I guess we're talking tenth's of a second or maybe full second at max so maybe it's not a big deal? Also the number itself doesn't tell much to a person who has no idea what to expect.
Quote:
Yes indeed. 32bit colors/1024 textures can be huge performance overhead for some older cards (like my GF2 GTS was) but 16bit could be entirely playable for example and doesnt even look that bad really. p.s. Thx for the kind comments/sticky Last edited by vari : 03-02-2004 at 19:36. |
#11
|
||
|
Funny thing looking at the results there, AA and AF seems to have no effect on the framerates. I guess this is because CPU is bottlenecking and the gfx card has virtually nothing to do
I ran the max test at 1600 resolution and then it started to slow things down halving the average frames. |
#12
|
||
|
Nice one Vari, thx
The good side from the "cpu-bottelneck" is that I can run LFS in 1600x.. with lot of AAA`s And that looks realy fine. Some suggestions for the next Benchmark (for S2 or whenever) -disappearing the chat line and end of Benchmark with finishing the lap. -min settings without wheel/arms min: Frames: 11365 - Time: 109844ms - Avg: 103.464 - Min: 71 - Max: 137 min, Wheel/arms off: Frames: 12635 - Time: 109500ms - Avg: 115.388 - Min: 76 - Max: 156 cu Mounthy |
#13
|
||
|
vari ...using 8x aa and 8x af here on the max settings really hurt my performance with my new fx 5600 ...this could either be a problem with aa/af in lfs or my card (runs great in other games)
|
#14
|
||
|
Yeah. What comes to AA, LFS just don't care about it. I could play the game @ 2048x1536x32 + 4xAA. I have a few screenshots too
__________________
LFS NAME SlamDunk | PLATE ... -.. | CONTROL Logitech M-S48a ball mouse + keyboard | FAVOURITE South City Long Rev | SlamDunk's Computer |
#15
|
||
|
This is a nice little test. I'll have to give it a go when I get home. It'll be nice to see a comparison chart of graphics cards and processor speeds/type. It may give LFS users an idea of what's important and who beats who. ATi versus Nvidia, Intel versus AMD, it would be interesting to see when we get some more results and a good variety in each catgory.
I did make a little frame rate test of my own a couple days ago, before I came across this. I left it pretty basic but went through a full range of Windows video settings and in-game texturing and resolution settings(anything short of disabling things/limiting draw distance). I ran it by myself and with a full grid of AI for comparison. Oddly, the in-game texture settings did pretty much nothing for framerates. Resolution affected it some. Windows AA, Antistrophic, and Texture settings varied the framerate some, about 2x at min. When I ran with a full grid of AI, it was another story. Basically reguardless of what I set for graphics, framerates were horrible and essentially unchanged. I just went with min observed framerates, not average or max. If you're not near the AI, graphics settings help, but packed in with them, nothing will save you, lol. I'll have to post my little results later. It might give an idea to another test, possibly a full grid test. I don't think you can run a spr though and test AI and their cpu usage affect on framerates. It would have to be in real time, and that's not perfect. Anyways, off to class.
__________________
Flying Squirrel
|
#16
|
||
|
i agree with that mvw2 for some reason the ai takes up a lot more cpu time then when im playing online with real ppl .. wierd :S
|
#17
|
||
|
Well with AI it's gotta calculate out where they're going etc., but with online players it's just receiving the antics of the other cars.
That didn't make any sense at all did it... |
#18
|
|||
|
Quote:
It's not weird at all. Like MGBoy says, online you receive info where the other cars are going, offline you have to calculate the driving off the AI cars as well.
__________________
The LFS Endurance League | Forum LFS Allround Rank: -0:36.140 (50 %) | +1:29.460 (100 %) GPL Rank: Working on it.... |
#19
|
||
|
Could someone write some decent instructions ... becase those are perthetic!
__________________
http://us.f1f.yahoofs.com/bc/2b420f...fnizGABQL01C134 |
#20
|
||
|
Thanks.
I'm gonna make new CFG's after the final version of patch E(5) is released. (It may change still so I won't do it now) |
#21
|
||
|
nice one .
get onto ti it shortly and let you know.
__________________
"When I die, I want to die like my grandfather who died peacefully in his sleep. - Not screaming like all the passengers in his car." - Emo Phillips |
Thread Tools | Search this Thread |
Display Modes | Rate This Thread |
|
|