Vanilla G400 Q3 results!

[Sven23]

doods systems are set up right.

I have a plain jane vanilla G400 not oc'd
Celeron at 458mhz 83mhz bus

Quake3 1024x768 32bit, everything on high

26.7FPS


tnt2s are getting 29!?!? 26.7 is pretty good for a so cold *slow* open gl, you guys with yer g400 maxes and piii cpus, your hardware isnt configured correctly?


let me know what up? Anyone getting scores like mine with the same setup?

loobygoat@earthlink.net

[Joelg]

Can you tell me how to run the timedemo under the latest 1.08 release?

Joel

[Maniac]

If your looking to actually play Q3 instead of just benchmarking it, I suggest turning off 32-bit textures and trilinear. Image quality remains very high and the frame rate jumps to 55fps at 1024X768 on my G400 O/C to 150/200...

[paulcs]

Joel,

To benchmark with demo 1, at the console enter

timedemo 1 [hit enter]
demo q3demo1 [hit enter again]

After the demo runs, you'll get bounced back to the main menu. To get your scores, bring down the console again.

Paul
paulcs@flashcom.net

[Joelg]

I am running a G400 32MB DH (o/ced to MAX speed 150/200) on an AMD K6-2/400. At 1024x768x32bit, with 32bit textures, trilinear filter, and all graphics options set to max I get the following:

q3demo1 - 21.0
q3demo2 - 20.6

I will say that I still do not fully trust these benchmarks, because I played Q3 online for a couple of hours this weekend and it sure seemed alot faster than these demo scores would indicate. Hell I even scored second place in one round with between 8 and 10 other ppl, and for me that is pretty damn good. I'm not the best player in the world, just ask some of the guys here. But for my machine setup I still think that that is pretty damn good.

Joel

[tombman]

Maniac , you say that you get with a g400 at 150/200 55fps (????!!!!) at 1024*768*32 bit in q3demo1.bsp with bilinear and 16 bit textures and all the other options at the highest setting ??
If this is really true ..then WOW !!!!!!
What kind of cpu do you have , what options have you turned on in bios and powerdesk ?
Please reply ....you know my g400max behaves like shit so i need your settings !

[Slartibartfast]

Yeah, the benchmarks aren't that hot. I installed Q3 1.08 on a machine with:

P-233MMX
64MB
10GB
Riva128 4MB PCI

And the demo was giving me like 5fps, 10 fps max in 400x300, but the game was VERY playable at 640x480... shows benchmarks aren't everything.

- Gurm... dammit... Slarty

------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.

[Slartibartfast]

BTW, the setup on my system (and a couple scores):

P3-450 (non-overclocked, soon to be a 500)
256MB PC100 CAS3
31GB HDD (14+17)
G400 Vanilla 32mb (non-overclocked)
MX300
Fireport 40 UWSCSI
SCSI DVD3 (Toshiba)
SCSI Scanner, IDE burner, etc. etc.
Primary HDD's on IDE

I get 29fps on demo1 at 1024x768 with everything on and turned up. I get 37fps on demo1 with annoying things off (sky turned down, extra-big textures turned off, bilinear instead of trilinear, etc.)

Demo2 is substantially slower (28-ish after tweaks).

And bear in mind, this is unoptimized, "crap" (according to many), beta OpenGL drivers... no P3 optimization, etc. etc.

And I didn't turn anything off before doing this, which means ICQ, AIM, MSM, HotSync, Go!Zilla, PowerDesk, and half a dozen other memory-resident thingies were running at the same time as this.

- Slarty



------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.

[Slartibartfast]

How's this:

1. Did a fresh install of Q3Test
2. Started the file q3test.exe
3. Used the menus to turn everything all the way up.

Happy?

- Slarty, who doesn't screw with his .cfg files... very often

------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.

[Slartibartfast]

I don't turn off sound. That would be stupid. Nobody ever plays without sound, so what purpose does it serve to benchmark without sound?

I turn the sky down for the second set of numbers, because it's a hog. Look at John Carmack's .plan, and you'll see that he has already removed 2 or 3 rendering passes from the sky (next test it'll be moot, apparently).

I don't turn off v-sync, either, since I don't like tearing and it doesn't make a really big difference right now with the G400.

k?

- Slarty

------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.

[Stil]

In order to measure your VIDEO card performance, it is necessary to get all other variables (i.e. sound, other applications running the background, etc.) out of the picture.

The way people should be benchmarking: 1. Delete your q3config.cfg.
2. Start q3 with the following command line:

quake3 +seta r_mode 6 +seta r_fullscreen 1 +seta r_swapinterval 0 +vid_restart +s_initsound 0 +snd_restart +timedemo 1 +demo q3demo1

Description:

+seta r_mode 6 = 1024x768. Change this number for a different resolution.

+seta r_fullscreen 1 = Duh.. fullscreen as opposed to windowed.

+seta r_swapinterval 0 = Turn off VSYNC

+vid_restart = Applies the video settings

+s_initsound 0 = Sound is off

+snd_restart = Applies the sound settings

+timedemo 1 = Timedemo mode

+demo q3demo1 = Run the demo

[Joel]

I bench them as I would play them. IMO the only true way.

Joel

[Slartibartfast]

Thank you Joel. I can't understand people who insist on benchmarking a system without their stuff running. Yes, I could probably squeeze high 40's out of my system instead of mid 30's if I bothered to optimize it for Q3. But guess what? I don't even PLAY the darn game more than once a week!

And I have the second-lowest CPU-utilizing sound card, which makes my machine FASTER than the average machine with sound on.

Don't start spewing command line nonsense. Put the game in. Start it up. Don't fiddle with things. Just play it. If it's good, it's good. If it's GREAT, it's GREAT. If it SUCKS, it SUCKS.

- Slarty

------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.

[Stil]

What I was saying is: The point of a benchmark is to have a number (in this case frames per second) that is meaningful compared across many systems.

When you start adding personal configuration options, you can't compare your number with anyone else's number because the test scenario is different.

If you just go by one person's configuration, you really aren't benchmarking the video card--youre benchmarking "Joel's personal configuration." This is a mistake many so-called professional reviewers (cough, Tom, cough) make.

In the end of course it only matters how it plays on your system, but in order to compare it to other cards, you need the same config.

[Slartibartfast]

Ok, this would all be fine if the benchmarking code in Q3 weren't still broken. I don't care what they say, it's waaaaaaay off.

I can run the tests later the way you'd like, but I'm still not messing with my config. I'll just turn off the sound and kill all the background tasks...

- Slarty

------------------
I used to be a guy named Gurm. I used to have a signature. But that's all gone now. Now, my name is Ash... and I am a slave.