Q3 new benchmarks with 5.25

[Allan2]

Here ya go folks. I use the following settings because they are the settings I WANT to play Q3 with, not the ones I actually use when I play.

My specs are PIII500@580
G400 DH @142core and 190mem
256MB PC100 Cas2 Ram
Abit BE6 motherboard
(a lot of other stuff that isn't so important)

1024x768
32 Bit Color
32 Bit Textures
Triple Buffering
Texture quality slider at MAX
Everything on in game options

no vsync Beta ICD Q3demo1 24.6
Q3demo2 28.2

vsync Beta ICD demo1 23.9
vsync demo2 27.0

no vsync non Beta ICD q3demo1 14.6
q3demo2 27.8

Vsync non Beta ICD q3demo1 14.5
q3demo2 26.4

[CoolBreeze]

Hi,

At the setting i play at with 5.25 and Beta ICD:

1024x768x32
32 bit color
16 bit tex
tex 1 notch down from the MAX
no V Sync
Sky is off
Bilinear
Everything else maxed

Demo Map 1 : 47 fps

Demo Map 2 : 50 fps


Quality settings:

1024x768x32
32 bit color
32 bit tex
Tex maxed
Bilinear
no V sync
no sky

Demo Map 1 :42 fps

Demo Map 2 :44.2 fps

G400 32 OEM Single Head @ 148.50/198

------------------
P3 500(not overclocked)
128 MB PC100 SDRAM
SE440BX-2 Motherboard by Intel
Seagate 8.4 gig HDD
Diamond MX300 Sound Card
AOpen 40x CD ROM
Ethernet 32 bit PCI Card
Motorola Cable Modem
Matrox G400 32 OEM Single Head @150
17" CRT Monitor
Altec Lansing 2 Speakers and Sub

----CoolBreeze aka KAST----

[Jihad]

Question
I am missing something because I cant find the option to turn off the sky; and didnt Carmack say that the sky in 1.08 doesn't eat so much CPU time as before?

[Kx30]

How is it possible that the first one is getting ~25 fps whereas coolbrezze gets over 40 fps ?????
(1024*768*32 , 32 bit tex , all at max)

[Allan2]

He's not using the same settings I'm using is what I'd guess. That's about the framerates I get if I go down to 800x600. Hmmmmm.....

[paulcs]

I believe he did. I found the speed hit from enabling the sky to be negligible. People, however, do like cranking everything up and disabling the sky.

Not CoolBreeze. I think his settings are pretty sensible.

I'd like to make a proposal. I think we should come to a consensus on benchmarking with the Q3test, a middle ground. Not too fast, not too slow. Something that wouldn't be too slow for online play, but at the same time, wouldn't ruin the eye candy.

Here's what I propose:

Normal settings at the following resolutions and color depths:

800x600 - 16 bit
800x600 - 32 bit
1024x768 - 16 bit
1024x768 - 32 bit

Sound on, sky enabled, vsync off.

Leave the sliders and game settings at their default settings.

This limits variables somewhat (refresh rate and, of course, the settings themselves), and I think it's a reasonable middle ground. Many people just don't have the processor power to max all the settings and get a reasonable frame rate. We post our numbers, at least in part, to give people some idea of what to expect.

Right now, I think we're seeing a "tower of bable" scenario, with everyone using different settings: some playable, some not.

Paul
paulcs@flashcom.net

[Kx30]

Ok, coolbreeze : since your scores are the highest i suggest you make this test :
1.08 , demo1 , 1024*768*32 , 32 bit tex ,
tex detail FULL , BILINEAR , VSYNC OFF ,
SKY ON (no speedhit since 1.08 !) .
(everything else ON)

Please tell us what registry settings you used ! (force agp? , triple? , vsync ? , powerdesk 32 bit Z ? .....etc)

THX

[Allan2]

Paulcs,
Like I said the reason I chose those settings is because that's the way I WANT to play Q3. Also since multitexturing is not enabled in the current version of Q3Test I suspect a great number of G400 owners will be able to run at 1024x768 with all the eye candy enabled anyway.

I assume most people agree that the purpose of benchmarking is to compare how well your system performs compared to other systems.

If everyone posts different settings it's like comparing apples and oranges. Pretty useless stuff. If you care to compare with me use the same settings I did. If you don't want to compare, post your benchmarks using whatever the hell Q3 settings you want

[Snake-Eyes]

Good post Allan2. That's the same reason I use those settings for benching the drivers with Q3, too. The only thing I'd recommend to help everyone compare their performance (across systems) is to disable sound, since different soundcards use varying amounts of CPU time to do their thing. This is also the best way to judge the effect different processors have on a video card, short of swapping processors out of the same box.

BTW- Your results are REALLY close to mine (we're talking tenths of a frame per second differences). I'm using a P3-450 (o/c to 558) with a G400Max and 256MB Ram.

[SwAmPlAdY]

I agree with Paul. If everyone is using different game settings, then comparing all of our results is pretty useless.

Is it really that hard for all of us to have 2 cfg files? 1 for our own settings and another one for comparing?

Sheeeeeeeesh

SwAmPy

[Maniac]

Wow where did this new driver release come from? Anyhow my fps went from 38.5 to 40.8
this is at 1024X768 32-bit, 16-bit textures bilinear, texture quality 2/3
vsync off
sky off
light flares off
everything else on...(btw these settings are best for great image quality as well as speed)
C464A G400@150/200

[Liquid Snake]

funny, allan, i had the EXACT same scores as you did with the beta ICD. But my scores were much higher with the non-beta ICD.

[CoolBreeze]

Kx30,Tombman,Deprived.....,

I ran your test at the settings you gave:

1024x768x32
32 color
32tex
MAX tex
Sky enabled
No V sync
bilinear
16z
PD 5.25

Q3 test demo map 1 : 40.2 fps....

this is with forced agp x 1....and i must note...

I have found that enabling sky DOES give ME a ~2fps drop even with 1.08. Also...i have the NEW MGA Tweak...and when i have it running in the background (start programs minimized in the taskbar)...it helps with the framerate...for some reason. Make sure that it is running in the background then try it out. I still get the frame rates i have posted...it is no lie...50 fps on map 2....in 1024x768x32...ill try and post with some more settings 2morrow....as for now, im gonna sleep. Thanx for your time...good luck with config yer rig man...

------------------
P3 500(not overclocked)
128 MB PC100 SDRAM
SE440BX-2 Motherboard by Intel
Seagate 8.4 gig HDD
Diamond MX300 Sound Card
AOpen 40x CD ROM
Ethernet 32 bit PCI Card
Motorola Cable Modem
Matrox G400 32 OEM Single Head @150
17" CRT Monitor
Altec Lansing 2 Speakers and Sub

----CoolBreeze aka KAST----

[CoolBreeze]

and i had triple buffering....

[Alex]

Just to share my results:

Q3test 1.08 settings: (default except for the following)
1024 x 768
color depth 32
texture detail set to highest
texture quality 32

5.21
q3demo1: 19.9 q3demo2: 37.6

5.25
q3demo1: 19.9 q3demo2: 37.6

5.25 + beta ICD
q3demo1: 34.6 q3demo2: 39.8

Driver 5.25 and beta ICD combination makes playing Q3test 1.08 acceptable to me. Without the beta ICD, there are certain parts of q3demo1 (when there are lots of firing) looked like they were in slow motion. I know that slow motion behavior is not observed on TNT2U. I am curious what is in the beta ICD that removes that slow motion behavior.

My 2 cents.

---------------
P3-500 (OC to 560)
Asus P2B-S
256MB SDRAM
G400Max retail
SB Live
18.8 GB WD SCSI U2W drive
SCSI DVD and CDRW
...