I finally try the Radeons 9800XT, 9600XT

[congo]

Yup, I finally wore my freind down, and he brought not only the ATI Radeon 9800XT around for me to test, but a 9600XT as well.

He had already benchmarked the cards on his machine and his son's as well. Both were AMD system's slightly higher in spec than mine. We achieved better scores on my PC for some reason.

My spec: XP2600+ t/bred CPU, 1024mb PC2700 DDR, 333mhz FSB speed.

We formatted a Hard Drive, set up a fresh Win XP installation, tweaked XP a little to my liking and away we went!

The Radeon's were set up the same with the driver.

The GF4 TI4200 was set at the nearest equivalent Nvidia settings.

Catalyst 4.2 drivers for Radeon cards,
Forceware 53.06 drivers for the GF4 TI4200

In 3D Mark 2001, quality settings, Antialiasing and Anisotropic filtering off, 1024x768x32 res.


9800XT = 15,746 marks

9600XT = 11,238 marks

TI4200 = 11,604 marks


In FS2004:
All the cards were used in the Seattle default flight with scattered cloud, full settings - less ground scenery shadows and water effects. Antialiasing and anisotropic filtering on, 1024x768x32

Frankly, the 9600XT, (tried first) was a little disappointing, not offering much more than my current GF4 TI4200. It was 1 or 2 FPS ahead of the TI4200 in flying around Seattle with some cloud, in some situations.
The R9600XT was 3 or 4 FPS better than the TI4200 in dense cloud, which makes it a clear winner there.

The R9800XT was pretty impressive, not by outstanding frame rates (from 20 fps to 45 fps), but by it's ability to clearly mipmap distant objects, (no more blurry distant textures!) and it's good performance in dense scenery and heavy cloud.

The 9800XT thrived on the higher settings and resolutions when we upped the anty. A very nice Video Card indeed.

[bm]

I'm glad to see you found a difference! My 9800XT is doing amazingly well now, to begin with I thought something was wrong but it seems to have settled down since I installed the Cat 4.2's.
My 3DMark score was quite low to begin with - but then I found out you don't run the test with 6xAA and 16xAF! I managed to get a score of 6539 which I think is pretty good! (following futuremark's instructions)

So - did he go home with the 9800XT?

[congo]

Yes, he took it back home with him.......  :P

He's finishing off all the video work he's in the middle of before he installs it in his rig.

So, back to my TI4200 and the blurry distance in FS2004.  That's the trouble when you see something nicer, you want it then.  ;)

[Jared]

yeah, make syou jealous...

[MichaelH]

"9800XT = 15,746 marks"

Drool...

thanks for posting all that Congo...

[Daz]

has anyone tested the fx5950 ultra against the 9800xt?? would be interesting to see which performs better

[Clownloachlover]

has anyone tested the fx5950 ultra against the 9800xt?? would be interesting to see which performs better

Jan. 2004 PCGamer

3dMark03                      9800XT        6,070
                                            FX5950       5,333

3dMark2001 SE             9800XT       17,182
                                            FX5950       15,885

Aquamark3    Score       9800XT       44,451
                                            FX5950       43,750
                     GFX          9800XT        6,027
                                            FX5950        5,859

Quake III   1280x1024   9800XT      266/164 fps
                                            FX5950      278/180 fps

UT2003      1024x960     9800XT      186/105 fps
                                            FX5950      192/98  fps

Splinter Cell 1280x1024  9800XT      38 fps
                                            FX5950       41 fps

Halo             1280x960    9800XT      37 fps
                                             FX5950       37 fps


#/#    1st is no AA or AF 2nd is 4xAA and 8xAF  
UT2003 all settings maxed
Splinter Cell and Halo all settings maxed.


Also mentioned was until DX9.0 games come out(Doom3,HL2) difficult to benchmark in a meaningful way.

After about 2 months of going back and forth this past weekend I upgraded from a TI4600 to a ATI 9800XT. I don't know how a FX5900/5950 does, but I can relate a few things that I saw.

My specs are:
P4 2.8
1 gig 2700 DDR
Ante 430 watt
ATI 9800XT

With FS2004. I was at 1152x864 32 bit 2AA 4AF with my TI4600 getting 25 fps. Now I run at 1280x1024 with 4xAA and 16xAF with Truform enabled and get 30 fps about 96% of the time with all settings maxed. The lowest it gets is about 24 fps with thunderstorms.

Over the past few years I have heard that Nvidia goes with fps over eye-candy and ATI goes for eye-candy over fps and the best example I saw was with Nascar Racing 2003. With my TI4600 I ran at 1024x768 2xAA and 4xAF with about 50% eye-candy and was getting 30-40 fps. When I put the 9800xt in and ran the game with the exact settings I only got 30-38 fps.

Made me a little mad, so I started upping everything (eye-candy,AA,AF, resolution). The fps never budged. In fact the higher I went the fps actually went up. When all was said and done, I now run Nascar at 1600x1200x32 4xAA 16xAF Truform on, all the eye-candy like grandstands, campers in the infield,  and now I get solid 35-45 fps about 50% of the time and the other 50% is around 60 fps.

Clownloachlover