Geforce3 Ti200 smokes a Radeon 8500?

[rogue1979]

Ok, I knew the title was an attention getter, but here it is. Let me clarify first, I am not on crack, not drunk, not insane, and I don't have a preference for Nvidia or ATI. I have a Radeon 9000 running 300/250, I just until recently had a Radeon 8500 LE @ 300/315, and my father has a retail Radeon running stock 275/275 on a similar system. I just got a Gainward Geforce3 Ti 450 Powerpack for $69 shipped running 230/530. Using Win98SE, 30.82 dets and NVmax version 3.00.0072 I am getting amazing framerates. Something seems weird, but I have triple checked the results. 8300 in 3Dmark 2001 with a 1800MHz XP is close to normal, but the gaming framerates are outrageous.

Running either 1280 x 1024 x 32-bit at 2X FSAA or 1600 x 1200 x 32-bit no FSAA, it is stomping my old Radeon 8500 LE scores by 25-30fps in all games. I used 16-tap anisotropic filtering and high details in all tests. The Radeon 9000 pales in comparison and even my fathers 8500 retail is getting spanked. I verified visually that FSAA was on when used, and anisotropic is indeed on. Somethin' ain't right here, but I am not complaining. In Unreal Tournament I am getting an average of 90 fps at 1280 x 1024 x 32-bit 2X, and with FSAA turned off it averages 175. My Radeon 8500 LE @300/315 managed only 60fps/130fps at the same settings, using the first Catalyst driver release. All other games pretty much give the same results, in Janes WW2 at 1600 x 1200 no FSAA the Gainward rocked 75fps average and the Radeon 8500 LE hit 45fps, with the Radeon 9000 hitting 30fps. Still not wanting to believe the results, I tried 4 different dets and 3 different NVmax versions. Only the 30.82 in conjunction with 3.00.0072 NVmax produced the same results.

My daughers' machine is a Celeron 1100 on a 140MHz fsb running a Geforce Pro at 220/460. She was playing Unreal Tournament at 1024 x 768 32-bit no FSAA and getting a respectable 60fps average. After swapping to the same 30.82/NVmax combo she is now running 1024 x 768 32-bit at 2XFSAA and she is belting out 70fps now. I know these results sound amazing, but unless I totally lost my mind it seems this driver combo is some kind of magic potion. Someone with a Geforce4 Ti give it a try and see what happens.

 

[Vadatajs]

so, what's your point?

 

[jiffylube1024]

Originally posted by: rogue1979
Ok, I knew the title was an attention getter, but here it is. Let me clarify first, I am not on crack, not drunk, not insane, and I don't have a preference for Nvidia or ATI. I have a Radeon 9000 running 300/250, I just until recently had a Radeon 8500 LE @ 300/315, and my father has a retail Radeon running stock 275/275 on a similar system. I just got a Gainward Geforce3 Ti 450 Powerpack for $69 shipped running 230/530. Using Win98SE, 30.82 dets and NVmax version 3.00.0072 I am getting amazing framerates. Something seems weird, but I have triple checked the results. 8300 in 3Dmark 2001 with a 1800MHz XP is close to normal, but the gaming framerates are outrageous.

Running either 1280 x 1024 x 32-bit at 2X FSAA or 1600 x 1200 x 32-bit no FSAA, it is stomping my old Radeon 8500 LE scores by 25-30fps in all games. I used 16-tap anisotropic filtering and high details in all tests. The Radeon 9000 pales in comparison and even my fathers 8500 retail is getting spanked. I verified visually that FSAA was on when used, and anisotropic is indeed on. Somethin' ain't right here, but I am not complaining. In Unreal Tournament I am getting an average of 90 fps at 1280 x 1024 x 32-bit 2X, and with FSAA turned off it averages 175. My Radeon 8500 LE @300/315 managed only 60fps/130fps at the same settings, using the first Catalyst driver release. All other games pretty much give the same results, in Janes WW2 at 1600 x 1200 no FSAA the Gainward rocked 75fps average and the Radeon 8500 LE hit 45fps, with the Radeon 9000 hitting 30fps. Still not wanting to believe the results, I tried 4 different dets and 3 different NVmax versions. Only the 30.82 in conjunction with 3.00.0072 NVmax produced the same results.

My daughers' machine is a Celeron 1100 on a 140MHz fsb running a Geforce Pro at 220/460. She was playing Unreal Tournament at 1024 x 768 32-bit no FSAA and getting a respectable 60fps average. After swapping to the same 30.82/NVmax combo she is now running 1024 x 768 32-bit at 2XFSAA and she is belting out 70fps now. I know these results sound amazing, but unless I totally lost my mind it seems this driver combo is some kind of magic potion. Someone with a Geforce4 Ti give it a try and see what happens.

That's all I can really think of; 8500 should be neck and neck with a GF3.

 

[Kaiser__Sose]

what nvmax settings did u tweak?

 

[rogue1979]

My point is according to many members on this forum the catalyst drivers are giving the Radeon 8500 near Geforce4 Ti 4200 performance. The latest benchmarks seem to support this, I say hogwash! Bring on the flames....

I set the AGP to 4X, fastwrites and side-band addressing on, but the 30.82 doesn't support fastwrites. Anisotropic to 16-tap, geforce accelerated lines on, v-synch off, force S3TC v3 compression on, force multi-texture on, direct3D W-buffer enabled, fast MipMap filtering on, mostly leaving the other settings the way they came.

 

[shags]

I don't know if this will help or not. But I have a Sound Blaster Live 5.1. I was getting around 25-30 fps with UT2003, then I heard that there was problems that even the newest Sound Blaster drivers could not fix. The fix was to put it in software sound instead of hardware. I did this and my framerate went from 25-30 to over 60fps constant... More than doubling my former framerate. You should try this and see if it works.

 

[alexruiz]

Are your games OpenGL, directX 7.0 or directX 8.0????

In most of the legacy games (directx 7.0 and older) even a Geforce2 can beat a rad8500......

I have seen the opposite. My relatives have an AXP 1800+/radeon 8500 retail/kt266a and I have an AXP1700+/ti200/sis735..... I am using detonator 30.82 in w2k, they are using catalyst 7.76.......

The radeon system crushes the ti200 in 3dmark 2001, and it totally clobbers the ti200 running Serious Sam...... Strange

 

[OvErHeAtInG]

Let's not forget, he's running 230/530. That's a serious overclock. Stock is what, 175/400? I'm running 210/470 currently on my Ti200.

EDIT Wait, Rogue... Did I read correctly, you said you were using the first catalyst release to benchmark? If I'm not mistaken, the catalysts have gotten WAY faster over time, correct?

 

[rogue1979]

It was the first catalyst for the 8500 LE, and the second one for the 9000 (9051 amd 9062 I think). There wasn't much difference in speed. Some of the games are open GL, some directX.

 

[Showtime]

I get similar marks with my ti200
-show

 

[spanky]

are the systems configured the same?
from what i heard... the 8500 is a little faster then the ti200.

 

[OvErHeAtInG]

This is devastating to me... I just ordered a 8500LE, for a second rig... I was expecting it to wup my Ti200... aghh

Rogue, I want to try to imitate your settings.. right now I'm running the 40.41 and I get 6706 in 3DMark. Have you tried the 40.41?

I don't know anything about NVMax. Where can I get it? I think I used something like "NVTweak" before, I had trouble with it so now I'm just using Coolbits.

 

[Tates]

Happy for you Rogue.

You are the exception and not the rule.

Using the latest drivers would result in a more objective comparison.

BTW, who plays UT anymore when there's UT2003?

Let's see how the GF3 performs with DirectX 8 optimized games. Different results most assuredly.

OvErHeAtInG, you don't have anything to worry about

 

[rogue1979]

I compared the Radeon 9000 @300/250 using the latest catalyst driver. Unfortunately, I don't have the Radeon 8500LE @300/315 anymore, so that was with the earlier catalyst driver. But I have used both catalyst drivers with the 9000 and I didn't see much of a difference in speed, although visual quality that was already excellent got even better.

The 9000 has a slight edge in 2D quality over the G3 Ti450, but in 3D it is pretty equal, although both are micro clear on a 1280 x 1024 desktop.
Since the 8500 is supposed to be keeping up with a Geforce4 Ti 4200 now, I thought comparing the 9000 with a G3 Ti200 would be close. That's why I was hoping someone with a G4 would try this driver combo and see if it yields big performance gains like I have realized on the G3 and G2 already.

I haven't tried the 40xx series yet, 30.82 and Win98SE with NVMax 3.00.0072 is the ticket. I first downloaded this version of NVmax over a year ago, so it might be hard to find now.

Tate, how did I know that you had a Radeon 8500 even before I checked your rig?
I don't have a preference for ATI or Nvidia, remember I am still running both. I favor what gives me the most performance for the least money, doesn't matter who makes it. If the new Trident XP4 or the SiS Xabre really start kicking butt and the drivers are mature, than I will go for that. Looks like the Radeon 9000 or 9000 Pro are the best bang for the buck for now. 8500 LE is also an excellent deal. The Geforce4 Ti 4200 is still a bit pricier than the Radeons, so if we believe the latest reviews the Radeons are a much better purchase. But if the G4 Ti4200 can show the improvement I am talking about, that even at $119 it might be a much more powerful solution than anyone thought. If the driver combo improvements have the same results as on the G2/G3 than it will smoke my G3 hand down, as well as any Radeon 9000 or 8500.