View Full Version : UT2003 benchmarks

14th September 2002, 12:58
Well the demo is out, new drivers are out, so might as well see how the Parhelia stacks up to the competition. I ran UT2003Demo/System/Benchmark.exe and these were the results (w. GigaColor and Trilinear filtering enabled all the way)

1024x768 noAA: 113.557869

1024x768 FAA: 65.161781

1280x960 noAA: 78.214691

1280x960 FAA: 45.915192

14th September 2002, 14:54
Ok, here are my benchmarks:
1024x768, no FAA, Trilinear Filtering:

1024x768, no FAA, Aniso Filtering:

Unfortunatly I can't offer you any 16xFAA benchies yet, please have a look at my "UT2003 problems" thread in the gaming forum.

About the look and feel:
In the benchmark both AF and FAA makes the game a lot slower in flyby mode, you can see it above with Trilinear vs. Aniso, AF is almost 30fps slower there. In bot match mode the performance lost is much slower, that goes for FAA and AF, you can see it above that AF is just 2,7fps slower than Trilinear.
The game itself plays good, I am playing it in 1024x768, 16xFAA, Trilinear. I just played a 16 player botmatch on the bombing run map, and it ran ok, not really fast but playable.

14th September 2002, 15:10
Well here's the results with my R7500.

1024x768x32 no AA no Aniso

Flyby: 67,9 fps
Botmatch: 34,3 fps

Athlon 1,47GHz, 512MB SDRAM, R7500.

14th September 2002, 15:29
Where can I get the demo and how do I benchmark it?

14th September 2002, 15:33
Found it! :)

14th September 2002, 15:38
To benchmark it, just go to the "\system" folder in the UT2003 directory and selsct benchmark.exe :)

14th September 2002, 18:15
Resolution 1024-768
16x FAA on
Trilinear on
Flyby: 59,016628
Botmatch: 25.009186

Resolution 1024-768
16x FAA off
Trilinear on
Flyby: 69.636223
Botmatch: 26.902164

Resolution 800x600
16 FAA on
Trilinear on
Flyby: 68.342918
Botmatch: 26.769066

Resolution 800x600
16 FAA off
Trilinear on
Flyby: 73.470650
Botmatch: 27.882294

Looks like i need a CPU upgrade.

14th September 2002, 18:19

14th September 2002, 18:37
That's weird Guru, you have a huge gap between the two modes, 20-30fps more than everyone else...what CPU do you have?

14th September 2002, 18:55
Since this is the proper forum I'll put my GeForce 3 TI200@240/500 results in:

800x600x32 trilinear: Flyby 91, Botmatch 32
800x600x32 with 8x anisotropic: Flyby 78, Botmatch 31
800x600x32 with 8x anisotropic and Quincunx AA: Flyby 54, Botmatch 30

1024x768x32 trilinear: Flyby 89, Botmatch 31
1024x768x32 with 8x anisotropic: Flyby 57, Botmatch 31
1024x768x32 with 8x anisotropic and Quincunx AA: Flyby 36, Botmatch 26

14th September 2002, 19:13
P4 2.26Ghz+parhelia retail:

1024x768 noAA trilinear

1024x768 FAA trilinear

1024x768 FAA aniso

1024x768 FSAA aniso

14th September 2002, 19:41
Duh, i didn't used the new drivers in my previous test. I just found out that there where new ones. :)

after installing i got these results:

Resolution 1024-768
16x FAA on
Trilinear on
Flyby: 69.138535
Botmatch: 29.519949

Resolution 1024-768
16x FAA off
Trilinear on
Flyby: 79.487053
Botmatch: 29.578671

New drivers have a nice performance boost.

Strange that FAA makes no difference in Botmatch. Or does this mean that the botmatch is very CPU limited?

14th September 2002, 20:18
Originally posted by lecter
That's weird Guru, you have a huge gap between the two modes, 20-30fps more than everyone else...what CPU do you have?

PIV 1.8a @ 2GHz+

14th September 2002, 20:28
Botmatch is indeed most influenced by CPU speed most....

As for my benchies:

1024x768 FAA ANISO


1024x768 no FAA, default filtering

114, 52

14th September 2002, 22:45
For what it's worth, here's mine with a GF 4 4600 (detonator 40.41) on my T-Bird 800@ 960 or so.... :rolleyes:

800x600x32 no AA no aniso
flyby: 76.45
botmatch: 24.02

800x600x32 no AA 8x aniso
flyby: 76.43
botmatch: 24.05

800x600x32 AA 8x aniso
flyby: 69.86
botmatch: 22.80

With the level recommended number of bots, all of these seem to be playable - for me at least, although it starts to stutter a little with high action....

Time for a new processor.... :rolleyes:

I don't think you really want 1024x768 with those results, but lemme know and I'll do it.

That will be REAL nice if UPS pays for it... lol


14th September 2002, 22:56
Radeon 9700 Pro

10x7 no AA no AF
FLY: 152
BOT: 50

10x7 4xAA 16xAF
FLY: 95
BOT: 48

15th September 2002, 00:56
Dave ... quit it, you're making me itchy. :D How's that puppy working out for HDTV res'es and DVD playback?

15th September 2002, 01:18
Yeah, I'd like to know about the DVD playback and the TVOut (is there a full Overscan?) as well

(Sorry for being OT)

15th September 2002, 01:26
My results:
800x600 no AA no Aniso
flyby 65.55
botmatch 23.89

1024x768 no AA no Aniso
flyby 63.79
botmatch 23.27

Almost no difference between 800x600 and 1024x768
But i think i need a new CPU ;-)

15th September 2002, 02:50
Dave ... maybe you could start a new thread in GHW to discuss the video aspects of your new graphics card ... please. I guess I should get back to AVS and see what they're saying about this.

15th September 2002, 02:53
I figure this is the place to post it

Heres some bench
done on a radeon 8500 retail 64 meg
Just to compare (Im still waiting to upgrade i didnt made my mind yet between P and a 9700)

All these score are at 1024x768 32bit
on a P4northwood@1.8

Flyby: 111
Bot: 43

16x aniso/nofsaa
Flyby: 111
Bot: 43
no change at all which is surprising (suspicious ?)

16x/aniso and fsaax2
FB: 59..

16xaniso and 4x fsaa
FB: 24
bot: 19

I tried to O/c my comp to 2.25 ghz for the fun of it

noFsaa and no AF
FB: 125
bot: 54

16th September 2002, 04:02
Hmmm, just for fun, tonite i'll try to cook up some V5 and G550 benches...
if i get over 5fps i'll post them ;)

16th September 2002, 07:13
Without 16xFAA and AF @ 1024:
Flyby: 103,839737
Botmatch: 45,178829

With 16xFAA and AF @ 1024:
Flyby: 51,368389

Athlon XP 1666 Mhz @ 166 FSB
(5:2:1 tiler is used)
512MB-DDR-Ram @ 166 FSB
Parhelia - Bulk
Abit KX7
SB Live! 5.1

All Components are wirth the latest Drivers.

16th September 2002, 12:32
Just to warn you before you go to all the hastle installing those cards. The demo won't run on any 3Dfx or Matrox G-series cards.

16th September 2002, 14:05
Out of curiosity, are all of you running with EAX enabled? I am.

Xortam, Indiana, I will tell you more about the 9700 features as I test them. I have looked at DVD already but to be honest, I don't watch DVD's with my computer. I have the DVD player sctrictly for running software apps that are supplied on DVDs. If you tell me things to look out for, I'll be more than happy to look at that aspect. Also, I will be comparing it to my Home Theatre set up since that is what I use to watch DVDs.


16th September 2002, 15:55
Dave ... I don't want to pollute this thread too much with this side issue. I guess you don't have enough experience with the video to start a new thread in GHW/SW. The AVS folks seemed to prefer the DVD playback on the 9000 over ATI's previous gens. I was wondering about Gamma, T&Rs, color saturation, etc. I usually check out skin pores and such on facial close-ups to check on clarity. You'd probably prefer the HTPC output over your home deck unless its a progressive scan player. Anyway ... I'll just monitor AVS to get more feedback on this aspect of your card.

16th September 2002, 18:22
UT 2003 Demo


Flyby: 152.320267
Botmatch: 49.975044

Flyby: 152.435501
Botmatch: 49.959011

Flyby: 151.934494
Botmatch: 49.901531

Flyby: 142.084518
Botmatch: 49.754349

Flyby: 86.616623
Botmatch: 46.055515



16x Ansio

Flyby: 56.752377

6x AA

Flyby: 47.968014
Botmatch: 24.337336

16x Ansio & 6x AA

Flyby: 37.581856
Botmatch: 20.769403

16th September 2002, 18:24
This benchmark absolutely sucks. Not representive of real gameplay and it seems to overwrite UT2003.ini every time you run it with it's default settings. So no testing of Truform or other options in .ini.

We need a UT 2003 GUI benchmark and a real demo representive of gameplay.....oh and we need actualy gameplay in UT 2003 to give us an incentive to play it over UT. ;)

16th September 2002, 19:27
Took a peek inside of benchmark.exe....

So, since benchmark.exe is using the same .ini the only difference in major cause of problems with benchmarking with benchmark.exe is inability to adjust graphics options (unless you change the source .INI in this case MAXDETAIL.INI), according to the readme the difference in BOT battle (slower CPU=stupider AI?), inability of crappier graphics hardware to take advantage of options enabled in MAXDETAIL.INI.

So we need a Benchmark prog that can detect the host vid card.
Support all graphics options of said graphics card without having to tweak host video card settings. (Also the benchmark could automagically set the video card settings to default if necessary).
Automagically gray out unavailable options for crappy video cards.

BENCHMARK.EXE information



dm-asbestos?spectatoronly=true?numbots=12?quickstart= true -benchmark -seconds=77 -exec=..\Benchmark\Stuff\botmatchexec.txt ini=..\Benchmark\Stuff\MaxDetail.ini userini=..\Benchmark\Stuff\MaxDetailUser.ini -nosound -UPT -

dm-antalus?spectatoronly=true?numbots=12?quickstart=t rue -benchmark -seconds=77 -exec=..\Benchmark\Stuff\botmatchexec.txt ini=..\Benchmark\Stuff\MaxDetail.ini userini=..\Benchmark\Stuff\MaxDetailUser.ini -nosound -UPT -

dm-asbestos?game=engine.gameinfo -benchmark -seconds=70 -exec=..\Benchmark\Stuff\flybyexec.txt ini=..\Benchmark\Stuff\MaxDetail.ini userini=..\Benchmark\Stuff\MaxDetail.ini
User.ini -nosound -UPT -

dm-antalus?game=engine.gameinfo -benchmark -seconds=77 -exec=..\Benchmark\Stuff\flybyexec.txt ini=..\Benchmark\Stuff\MaxDetail.ini
userini=..\Benchmark\Stuff\MaxDetailUser.ini -nosound -UPT -

Files explanation

showhud - Removes HUD information. Text etc.etc.
Ship - removes demo version text

causeevent flyby - Some sort of scripted flyby sequence?
Ship - removes demo version text

MaxDetail.ini - Benchmark.exe uses these for the UT2003.ini to see benchmark information

MaxDetailUser.ini - Used to set control information for the benchmark?