[AlanApple]

Hi all,

I am testing my new GeForce 4 Ti which I run on a G4 800 with 640 ram. The low res framerates in Quake III were okay but the card maintained this on the high res setting. Does this sound normal.

Resolution 640 x 480

Test 1) 1260 frames, 23.0 seconds, 54.9 fps
Test 2) 1260, 22.8, 55.2 fps
Test 3) 1260, 21.5, 58.5 fps
Test 4) 1260, 21.5, 58.5 fps

Resolution 1280 x 1024

Test 1) 1260, 22.2, 56.9 fps

[PowerBroker]

I'd say yes, but I'm hesitant because I'm not entirely sure of the suggested performance. Perhaps I'll benchmark my own GeForce 4 Ti 4600, to see if I can reproduce your results.

[Jansar]

Originally posted by PowerBroker:
I'd say yes, but I'm hesitant because I'm not entirely sure of the suggested performance. Perhaps I'll benchmark my own GeForce 4 Ti 4600, to see if I can reproduce your results.

I'll do the same and post.

[Gul Banana]

Those results don't sound right to me.. I can get up to about 70 fps on a G4/400 with a Radeon 8500 with less RAM, so, you should be getting higher.

[AlanApple]

Hi all. Thanks for the input. Let me know how you get on.

The test I used before was a 'timedemo' but since then I have managed to get the fps on screen. Gives a much better picture. I put the screen res up to 1280x1024 and the results were very encouraging. While there was a low rate occasionally, such as when starting Quake III, it changed in a blink of an eye and the fps rate was firing up to 80/90. Good on such a high res. I even blinked into the 100 area.

Feeling much happier. I bought the card off Apple but when I asked them what i should be expecting they were not able to tell me anything. The technician did try his best and said if there was no distortion or obvious problems he felt it was okay. This is fair enough I suppose. But there is no access to support in this area for nVidia so there is a general lack of tech support. Not good for such an expensive card.

[Boochie]

Your latter results seem reasonable. I have the same card with a G4/867, and I can run Jedi Knight II at 1280x1024 with medium detail at refresh rates in the 70-90 range. They drop off when there are a lot of bots on the screen, but that's almost entirely due to CPU loading.

When I'm playing mutiplayer JKII (saber duels), frame rates stay consistently in the 70-90 range, and rarely drop off.

[AlanApple]

Thanks Boochie.

I am using a G4 800 with 640 ram so it sounds like we are getting similar rates. Like you I get some changes but only for a sec.

This is still the top of the range card - still only a bto on the new Macs - so I would imagine their is a lot of latent performance. I reckon that if i add more ram I will see better performance again.

Still not sure why the rates using the on screen fps counter were different from the timedemo.

I was talking to a graphics designer recently and he says the human eye can only detect up to 60 fps so after that is is academic apparently.

Has anyone played Sacrifice?

[bborofka]

Timedemos usually give lower fps than in-game fps. I think it's better too because then everyone can use the same standard, where as some people give subjective results like "the framerate dropped into the 40s when bots came on the screen."

What version of X and Q3A are you running? I'm running Q3A 1.3.2 in OS 10.2.3 on an eMac with 245MB RAM and a GeForce2MX, and I get 45.2fps on Normal settings when typing:

timedemo 1
demo four

[AlanApple]

I am using OS 10.2.3. on Quake III Arena version 1.32 with an Apple 17" LCD.

Below are the results from 'timedemo' running the different settings:

Graphic Settings: High Quality

GL Driver: Default
GL Extensions: On
Video Mode: 800x600
Colour Depth: 32 bit
Fullscreen: On
Lighting: Lightmap
Geometric detail: Medium
Texture Quality: 32 bit
Texture filter: Trilinear

1260 frames. 21.4 seconds: 59.0 fps
1260 frames. 21.1 seconds: 59.7 fps
1260 frames. 21.1 seconds: 59.8 fps

Graphic Settings: Normal

GL Driver: Default
GL Extensions: On
Video Mode: 640x480
Colour Depth: Default
Fullscreen: On
Lighting: Lightmap
Geometric detail: Medium
Texture Quality: Default
Texture filter: Bilinear

1260 frames. 21.1 seconds: 59.7 fps
1250 frames. 21 seconds: 59.9 fps

Graphic Settings: Custom

GL Driver: Default
GL Extensions: On
Video Mode: 1280x1024
Colour Depth: 32 bit
Fullscreen: On
Lighting: Lightmap
Geometric detail: High
Texture Detail: max
Texture Quality: 32 bit
Texture filter: Trilinear

1260 frames. 22.6 seconds: 55.7 fps
1260 frames. 22.6 seconds: 55.7 fps
1260 frames. 22.5 seconds: 56.0 fps

[timster]

The truth is that the GF4Ti is such an immensely powerful card that your G4 is outclassed.

Pretty much all the limits you're seeing on your framerates are CPU limits, not the limits of the card itself. A few months ago I bought a new dual 867 with the 4MX, and I had already accquired a 4Ti, also since I was migrating from a dual 450 with a flashed 8500, I took the opportunity to compare all three cards.

What I found out was that the 4MX, the 8500 and the GF4Ti didn't offer much variation for me unless the game could offer SMP. Case in point - (I'm unable to find the notes I had so this is from memory):

Quake III (SMP enabled)

at lower resolutions, the 4MX, 8500 and 4Ti posted nearly identical numbers, 110-130fps.

it wasn't until 1024x768 that the 4MX started to fall off, and then at 1280x1024, the 8500 started to dip off a bit while the 4Ti remained at 130fps - obviously not affected yet. My 18" LCD doesn't do 1600x1200 so I wasn't able to test that resolution.

Return to Castle Wolfenstein
(this game is not SMP enabled)

Every card here posted about 44-46fps. Regardless of resolution or video card. Dropping the texture qualities would increase the fps to about 52-54fps for all cards. The real surprise here was that the 4MX actually performed better than the 8500. Drivers, I guess. RtCW was just too cpu-heavy for the video cards to distinguish themselves.

So long story short - unless you've got something faster than a dual-1GHz G4 and run at really high resolutions.. you're not going to do any better than a 4MX on most games out there.

[AlanApple]

This is good to know as info on graphic cards does not seem to be easy to come by. No tech support line for nVidia - what's that all about?!

Did you get your fps count using a timedemo or was it an on screen pfs counter?

Cheers

AA

[rampant]

Originally posted by AlanApple:
Hi all,

I am testing my new GeForce 4 Ti which I run on a G4 800 with 640 ram. The low res framerates in Quake III were okay but the card maintained this on the high res setting. Does this sound normal.

Resolution 640 x 480

Test 1) 1260 frames, 23.0 seconds, 54.9 fps
Test 2) 1260, 22.8, 55.2 fps
Test 3) 1260, 21.5, 58.5 fps
Test 4) 1260, 21.5, 58.5 fps

Resolution 1280 x 1024

Test 1) 1260, 22.2, 56.9 fps

Those are patheticly low scores, sir. I think I know what your problem is.

Go to console and type "r_swapinterval 0" then "timedemo 1" and run the benchmark. It looks like you're being limited by your refresh rate (are you on a LCD?) For regular gameplay, however, keep "r_swapinterval" set to 1. Nobody likes torn frames.

[littlegreenspud]

On my Dual Gig QS 1.5G RAM I see around 200FPS at 1280x1024 with all details at max in Quake. But Quake is MP aware.

[tpxstyle]

i get 70-80 fps on my QS733 w/ radeon 7500 =X

maybe you should try the com_hunkmegs XXX
XXX= amount of memory allocated to the game
i think about half of your systemmemory would be good..

[AlanApple]

Thanks tpxstyle.

Did you get your fps count from a timedemo or the on screen fps counter?