UT2003 Benchmarks

[Thomas @ Kustom]

Hey guys,

I'd be interested to see your UT2003 Benchmark results.

For those that don't know how to run this here's how.

Download UT2003 Demo from here http://www.unrealtournament2003.com/downloads.php

Install, then go to the game directory, usually c:\ut2003, within that there is a folder called system, you'll see in there there's a benchmark tool. Before running this, go into the demo by running ut2003.exe and set everything to highest settings, I'm not sure if this varies depending on whether you have the full game or demo, the game may allow higher gfx settings.

After you have applied these and saved, exit the game again and run the benchmark, it will ask you for the res, so let's keep it to 1024x768.

After it has run 4 benchmarks(doesn't take long) it will give you 2 scores, the flyby and the botmatch.

Let's see some results with specs.

Flyby _____
Botmatch ___
CPU ___ FSB ___
Mobo ___
RAM ___(include make and speed)
GFX ___(Include make and model)

Will be interesting to see how much this differs from machine to machine. :)

I'll post mine tonight.

[conan]

Mine are, Flyby 151, Botmatch 69, with rig as in spec below with same OC on CPU & GPU.

On stock settings for both CPU & GPU I get Flyby 122, Botmatch 51.

BTW, this is one of the benchies I use when checking overclocks. The tests were done @1024x768 32bit.

[PhilD]

Flyby: 97.876900
Botmatch: 34.223824
CPU: 1.8@1.9 FSB 105
Mobo: Chaintech Apogee 9EJL1
RAM: Crucial 512Mb DDR266@279 2100(include make and speed)
GFX: Leadtek Geforce FX5600 256Mb VIVO(Include make and model)

Just started clocking this setup for the first time about 45mins ago so i'll edit when i've got it all peaked.

::edit:: bit of tweaking still at 1.9:
Flyby 111.919243
Botmatch: 39.565384

[slartiBardfarst]

Flyby _____181.835602
Botmatch ___79.335323
CPU ___2380mhz FSB ___170mhz
Mobo ___gigabyte
RAM ___768mb crucial
GFX ___9700 non-pro @ 410mhz core \ 620mhz mem

all that it it never got above 31 degree's, and it's very hot today :)

[Thomas @ Kustom]

It's interesting, it would seem that a high FSb results in good Bot match results...

[MetalStorm]

High FSB usually means its a fast computer though... clock speed does have a HUGE impact on FPS in Ut2k3, GPU also makes a lot of difference unless you have nothing to back it up, a R9700 pro will run the same speed on a 1.3GHz Celeron as a Ti 4200 for example.

I will post my scores in a minute...

Okay, I have run the benchy:

Flyby - 174
Botmatch - 76

The Flyby results I think are still limited by CPU speed, Botmatch definatly is

[Thomas @ Kustom]

It's frustrating,

I'm trying to help a valued customer to get the botmatch scores in particular up, but does anyone else ever experience slow performance in ctf-face3? he says his stutters, I'm trying to figure out what it is because his system is
p4p800
3ghz(800)
512 pc3200 memory(can't remember brand, but it's one of ours)
ge-force fx 5900 ultra 256mb(gainward)
audigy2

I can't remember much past that, it's windows XP.

For some reason he says loading a few levels before ctf-face3 makes it better, if anyone else has or had this problem and knows how to solve it please let me know. It's got to be something software because the system is up to it!

[conan]

I'm not sure but if it runs better when loading a few levels first is it possible that some common data or routines are being cached in the CPU/HDD?

[The Jez Man]

I ran it four times, twice with the stock CPU settings and twice with it OC'ed. Once with AA and AF off and the other time with em at 6xAA and 16xAF

Flyby: 216.6 || 112.7
Botmatch: 77 || 62.4
CPU: P4 3.0ghz FSB: 200
Mobo: Soyo P4I875P
RAM: 1gig Corsair XMS 3200 [400mhz]
GFX: Radeon 9800pro 128

Flyby: 228 || 112.7
Botmatch: 85.6 || 63.3
CPU: P4 3.3ghz FSB: 220
Mobo: Soyo P4I875P
RAM: 1gig Corsair XMS 3200 [@440mhz]
GFX: Radeon 9800pro 128

- Jez

[MetalStorm]

I would have said that the reason CTF - Face 3 is juddering is because it uses over 190mb of textures, and obviously unless you have a 256mb card then its going to need to reload them as it goes, but you say he has a 256mb card... so that doesnt make much sence the only thing I think it could be then is if he is using AA and AF as well, as they use quite a lot of memory and so will take the effective available memory down so that it has to reload textures occasionally...

I know that Face3 used to jerk once at the beginning when the map was loaded, becuase the HD was still caching the textures, that problem is now gone with my RAID array, so maybe he should get a faster HD? - He seems to have enough money as it is!!!

The only other thing I could think of is that it might be OS related, my set up runs fine, and its nothing in comparison - a reinstall usually helps :) Maybe suggest reinstalling on another partition or using a spare HD just to test to see if the OS is the problem?

[Thomas @ Kustom]

Hi Metal,
Thanks for the suggestions. His hard disk is a UDMA133 120Gb 7200rpm Maxtor DM9 8mb Cache, believe me, hard disk speed isn't the problem, ok it wont compair to raid, but it shouldn't need to be that level.

I think he does run a few options like vsync etc, not sure about AA, i think this was off.

I don't think it's fair to say that money isn't a problem, you don't know how long someone may have saved for something like this system.

I'm building a replica system of his today hopefully to sort the problem.

[MetalStorm]

Okay, sorry about the money bit.

I have thought about that problem some more and this is what I have come up with:

Either his HD needs a good defragment, considering the size of the map and the number of textures that could always be a problem, especially if it is very fragmented.

Video card drivers - if he had an ATI card before, they could be conflicting and disrupting if not propperly uninstalled, this could drop fps quite a bit.

AA and AF, the 5900 Ultra might be a good card, but its 8xAA sucks badly, a lot more than it should, it takes a huge performance knock from having it on, and it doesnt even look much better than 4xAA... I would suggest making sure they are off...

Game settings, the reason the game might be jerking in that map might be due to him having a lot of bots in game, bots really do eat up CPUs so maybe he was playing with 30 bots or something?

The last thing I think it could be is the OS, sometimes things unexplainable like what he is experiencing are fixed with a nice format and reinstall - though I did previously suggest trying his hardware out on a fresh install on another HD or partition so if that isnt the problem he doesnt have to spend ages putting all his stuff back on.

I hope some of that helps!

[The Jez Man]

The only thing I could suggest [and I'm a simple minded fella] is to update the mobo bios, it may be as simple as that especially if it is a new mobo [The good guys down at ASUS may have forgotten something in the orignal bios]. Admittedly he might not want to do it himself especially if there is a risk of getting it wrong.

- Jez

[dtsrules]

flybuy 172
botmatch 75

p4 2.8c @ 3.1
msi 875pe
2 x adata 512meg ddr500
9800 pro

aa and af on full
115
54

[Remo]

flyby 209:
bmatch 78:

all @ stock values

[Slamdog]

flyby 209
botmatch 66

p4 2.4b @ 2.880 with raddy 9700 pro ram, 2*512 twinmos matched pair pc3200

[Chris0288]

Flyby _____230
Botmatch ___107
CPU ___A64 3200 @ stock
Mobo ___Asus K8V Deluxe
RAM ___2x512mb Mushkin PC3500 Level 2
GFX ___Sapphire 9700pro @ stock

i was shocked when i ran the test the first time:confused:

[MetalStorm]

Why were you shocked???

[Chris0288]

i came from a xp2000 that got 130fps flyby and 50 oddfps with same graphics card, needless to say i knew these things were fats but THAT fast!

[Shocky]

Nice score Chris. If people havent figured it out yet the Unreal engine is CPU limited and as you can see by his results the A64 shines in Unreal engine based games. Do you play UT2003 chris or did you just download the demo to benchmark it?

234.91|160.61
84.80|80.11

NO AA/AA | 6xAA/16xAF

[bouncin]

my scores are

flyby 342
botmatch 109

hmmm thread is a bit old now but good to see the comparison between old and new