Sign in with
Sign up | Sign in
Your question

Catalyst 8500 vs. my GF3 original

Last response: in Graphics & Displays
Share
June 17, 2002 6:20:40 AM

Had a lan party this weekend and ran a Q3A demotest against a buddys 1800+ powered 8500 (retail) rig versus my 1700+ powered GF3 setup..

same demo test, same settings (1024x768x32).. results..
Him 130FPS
Me 140FPS

I know thats pretty close but my point is (besides I'm happy) that a 8500 with new drivers still can't beat my GF3. On paper the 8500 is supposed to be like a 4400!

Another friend there had a 4600, the 8500s real competition (both are each companys front runners right?) and we all know what happened there.
Geforce3 killer I say not!

Athlon XP 1700+,KT266A,Geforce3, Audigy.. 'nuff said.
June 17, 2002 7:26:29 AM

keep in mind that the result also depends havily on driver setting, ie, high performance will give u much higher framerate. with eveything on performance, i dont' see y the 8500 can't beat the GF3 ( is it a TI btw? )

another thing is that, both of your system specs should be identical, ie, same amount of ram, same mobo, same setting in bios ...etc and nothing should be running at the window's background while the benchmark are been carried out.at the end of the day there are just too many things that needs 2 be controlled inorderto get a fare result.
that's y @ toms, they only use one mobo but differnt GPU + a clean window install before the tests are been carried out.
June 17, 2002 7:50:15 AM

both of your systems have too many variables in them to acuratly compare the two cards...
Quote:
Another friend there had a 4600, the 8500s real competition (both are each companys front runners right?) and we all know what happened there.

the R8500 was not designed to compete with the gf4 ti series (except for maybe the gf4 ti4200)...it was only designed to compete against the gf3 series (ti200, original, and ti500)..which it does pretty well...


:eek:  <b>Who fixed <font color=red>ATI</font color=red>'s leaky faucet??</b> :eek: 
Related resources
Can't find your answer ? Ask !
June 17, 2002 11:04:53 AM

Weird, 130FPS only? With a couple of minor tweaks, I get closer to 160FPS on my 1.33GHz Athlon. My R8500 is not overclocked.

The R8500 should almost always outperform the GF3 and even now, the Ti500. Otherwise, there is some kind of conflict or misconfiguration.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 17, 2002 11:35:22 AM

or maybe the setting is on balance or high quality
June 17, 2002 11:55:21 AM

Probably. In HQ, Aniso is on, which will take a few FPS toll on performance.

IMO, High Performance + Aniso seems to be the best setting, whether it be for OpenGL or DirectX. I'm not entirely sure if its actually working for OpenGL or not, though (High Performance mode, that is).

In DirectX, I can't tell the difference in IQ.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 17, 2002 7:51:54 PM

We both had anisotropic filtering turned off in our drivers, as far as high quality in Q3A, I was running high quality but i'm not sure he was.. I think I had my settings a bit higher as far as textures are concerned ect.

I figured turning aniso off would give me an advantage as the 8500 is supposed to be faster with it on.
But with high quality mode in Q3A on then I had aniso on you say? (or do you mean high quality mode for the radeon, which is what i'm assuming)

Our systems are not exact but I'd have to say he had the advantage if anyone did, 1800+ w/512MB PC2100 DDR versus my 1700+w/512MB PC2100 DDR..

mine isnt the ti series just the original GF3 (visiontek).

I dont want to cause trouble but I'd like people to post real world comparisons.

Remember this isnt the same demotest you guys are running, this is a custom one the lan party host had on his server, so he may have gotten 130fps in this particular one while you get 160 in another..

Athlon XP 1700+,KT266A,Geforce3, Audigy.. 'nuff said.
June 17, 2002 8:14:20 PM

Quote:

this is a custom one the lan party host had on his server, so he may have gotten 130fps in this particular one while you get 160 in another..

Wait, are you saying this demo was running off another system? Do you both have the same network card? If not then it's probably a network latency issue.

Seriously, with current drivers, the R8500 should outperform any GF3 in almost everything especially games that use the Quake 3-engine. With Quake 3-engine games, the R8500 matches and even outperforms the Ti4200.

What did you have the slider set at for OpenGL?

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 17, 2002 8:38:54 PM

No.. it was simply a custom demo of a fragfest in the bouncy map, downloaded from the server.. not running from the server. I'm insulted you think I'm that naive!

I dont know, I just saw what I saw.. as far as his driver settings he said he had them optimized (performance oriented settings) but i didnt set them for him.
I always run aniso8x and got 80fps in the same test with that on but turned it off and got the 140.

I believe you but there were no strange circumstances in this case that I knew of and this is the only real world test I've seen with my own eyes.

Athlon XP 1700+,KT266A,Geforce3, Audigy.. 'nuff said.
June 18, 2002 12:06:44 AM

O, so is that a challenge? Will ya compete w/ my Quadro DCC then in Spec View Perf?

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 18, 2002 5:06:26 PM

I have to ask, is his a retail running at 275/275? Or does he have an LE/OEM 8500 posably running at 250/250 or lower?

Also, when he upgrades his drivers, does he manualy delete old registry entries and files before installing the new ones? The first time I did that, I got an extra 5-10% boost from a driver upgrade.

"Meesa thinks that yousa gonna die" - Darth Darth Binks
June 19, 2002 1:22:32 AM

Quote:
Seriously, with current drivers, the R8500 should outperform any GF3 in almost everything especially games that use the Quake 3-engine. With Quake 3-engine games, the R8500 matches and even outperforms the Ti4200.


Not according to the benchmarks I have seen, including toms own.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 1:26:58 AM

yeah, I have to agree w/ ya Mat, Ati guys always say that, but I'm yet to see it outperform a Ti500!

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 1:30:38 AM

You're kidding right?

What's the deal with lampshades, I mean it's a lamp, why would you want a shade? :smile:
June 19, 2002 1:47:42 AM

No, actually I'm not.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 2:02:20 AM

Why is it that nVidia fans can't believe that ATI is doing wonders with each driver release in terms of performance. First a huge boost in OpenGL with the 6043 and now a huge DX boost with the Catalyst.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 2:06:06 AM

Can you say <i>bias</i>? Flamethrower and Matisaro look at the 6043 and say the R8500 is slow when that was 4 driver revisions ago! The Tom's VGA Charts are dated! Extremely so! The 6053, 6058 and 6071 had some impressive improvements. I'm seeing a 15% boost in D3D with the Catalyst in High Performance mode with absolutely no noticable loss in quality.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 2:06:35 AM

I am saying that the benchmarks I have seen dont impress me to the level of making the BOLD claim an 8500 will beat a 4400.

Catalyst....the review I read(and I old read one cause I dont care about graphics cards much anymore) showed a whopping 6% increase IN 3DMARK2K1, and commented that the gains didnt seem to transfer to common games.

And every ati head is always bragging about their drivers, when the cold hard truth is every review I have EVER SEEN from the 8500 shows it slightly edging out the ti500 or losing to the ti500, let alone the gf4 series.

And I cant understand ati fans and their constant quest to enbiggen the results of the 8500, making bogus claims it is better than the 4400, when it isnt even close, UNLESS you turn on 64 tap aniso to which I made my aa comment.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 2:09:09 AM

Quote:
Can you say bias? Flamethrower and Matisaro look at the 6043 and say the R8500 is slow when that was 4 driver revisions ago! The Tom's VGA Charts are dated! Extremely so! The 6053, 6058 and 6071 had some impressive improvements. I'm seeing a 15% boost in D3D with the Catalyst in High Performance mode with absolutely no noticable loss in quality.


Im not biased, im saying I havent seen any benchmark results which impress me at all, perhaps you should link to some, and high performance has to be doing SOMETHING, so I ask you, what is it doing, furthermore, given atis history (quake 3 incident) how do you KNOW its not doing anything to lower quality.

If you want to make the claim the 8500 is BETTER than the 4400, and you did, then you had better have some benchmark links to back that statement up, and if you give me 64 tap 3dmark2k1 benchmarks I will laugh at you.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 2:09:15 AM

When did I say it was better than the Ti4400? I was talking about the Gigabyte R8500XT!

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 2:10:01 AM

And I am asking for proof of that statement.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 2:10:20 AM

Well it's a tough call. People like nvidia because they have OK drivers while ATi just doesn't seem to be reliable all of the time, and people can easily become bitter about it (I know I can). However, everyone here must admit that no 3D card drivers are really all that good. You must manually uninstall them all and delete all registry entries, which is a hassle, before installing new ones. Same is true for nvidia, ATi, all of them. Good thing there's detonator destroyer but too bad there is nothing for ATi.

And Kinney, I think people are being kind of patronizing towards you because benchmarks must always be done fairly and using the same exact setup each time. The hard drive fragmentation may affect framerates. The motherboard obviously will. BIOS settings will. Memory manufacturer and brand will. In-game settings will. How you installed the video card drivers will affect it too. So will the operating system and what drivers you have installed for it. So will the drivers you have installed for your mainboard. There are just too many factors to take into account. Personally I think nvidia is an okay company, but I myself am biased towards ATi because I think they have better image quality. But even though I am biased I think that that is an unfair test because you used two completely different systems.

Censorship makes us so much more creative.
June 19, 2002 2:11:34 AM

Quote:

Catalyst....the review I read(and I old read one cause I dont care about graphics cards much anymore) showed a whopping 6% increase IN 3DMARK2K1, and commented that the gains didnt seem to transfer to common games.

At <b>default</b> settings which provide minimal improvements. Set to high-performance a 15% boost is possible with no noticable quality loss.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 2:15:38 AM

Quote:
At default settings which provide minimal improvements. Set to high-performance a 15% boost is possible with no noticable quality loss.


Look, at default, what does high performance change, it has to change something the quality HAS to go down to get the extra performance, SOMETHING needed not to be shown on high performance, just cause YOU cant notice a quality difference doesnt mean there ISNT one.


Furthermore, if there truely was 0 impact, 100% NO quality loss, then why not have high performance set as default, why even have a choice, could it be that there IS IN FACT a quality loss, and that quality loss makes said gain invalid in comparison to gf4 which does not sacrifise quality for frame rate.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 2:18:05 AM

The difference is that it compresses the textures further and it uses slightly lower-resolution mipmaps.

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 2:29:24 AM

And thats a quality difference, you use subjective terms like slightly, the fact is the high performance drivers change the quality settings, and even though YOU claim you cannot tell, it is a different setting, and thats no more fair than running the 8500@640x480 and the ti500@1070x768.


You have to use normal, not high performance to compare videocards, just like you cant lower quake 3's detail levels without saying so to boost your performance.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 3:11:25 AM

funny thing cake, I've never uninstalled an Nv driver in my life! I've just installed the update over it and never had a problem....same w/ all my freinds who own nv, and u know what, ati does make bad drivers. I have seen this peronslaly now on 3 separate occasions, w/ all drivers from ati.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 3:13:19 AM

The retail, 275/275, he wouldnt have the oem version for anything.

Athlon XP 1700+,KT266A,Geforce3, Audigy.. 'nuff said.
June 19, 2002 3:34:52 AM

Well Tom's reveiws are very very very very very very old and very very very unupdated. ATI has since had newer versions of the r8500 out and 5-6 driver updates since so they don't count.
June 19, 2002 3:47:42 AM

So guys you wanted proof well 3dgameman has A VIDEO REVIEW (SEEING is BELIEVING) of the a radeon 8500 128 mb running catalyst drivers. A quick summary from 3dgameman (website 3dgameman.com)follows as this------------------The ATI All In Wonder Radeon 8500 128MB Video Card is a step up from the 8500DV for gamers offering better results in 3D games. As a matter of fact, using ATI's latest CATALYSTâ„¢ 02.1 drivers, gaming performance is on par with nVidia's Geforce 4 Ti4200 based Video Cards. Also, this card has dual display support, S-video in/out, built-in TV tuner, remote control, and excellent 2D/3D performance packaged in one card.


It is on par with the ti 4200. Well guys ESPECIALLY YOU FLAMETHROWER YOU WANTED PROOF AND I GAVE YOU A VIDEO what do you have to say now. By the way don't try downloading the high quality video unless you have a broadband connection cause it is more than 30 mb.
June 19, 2002 3:50:26 AM

Please do give me the link (I think 19MB/s is broadband). I never trsut summaries, b/c they are so easily tainted to bend the facts.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 3:50:55 AM

So whats your point cause I can't see it.
June 19, 2002 3:51:24 AM

did u read the post?

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 9:10:31 AM

I have swapped video cards around a bit, using both ATi and nvidia cards. When uninstalling nvidia drivers, there is still junk left around for me, such as nview. The drivers are fully functional, and only in rare cases do uninstallation issues affect actual performance.

Summaries never do tell the truth, usually just a twisted form of it. Sometimes summaries take the form of an excited reviewer who wants so badly for his new review sample to beat whatever is the status quo that he writes a dazzling summary and people believe it as fact. Fortunately tomshardware doesn't do this, and anandtech definitely doesn't do this.<P ID="edit"><FONT SIZE=-1><EM>Edited by cakecake on 06/19/02 05:20 AM.</EM></FONT></P>
June 19, 2002 4:08:56 PM

Ok, lemme get this across right now- I really could care less about 3d mark- ok, ati shines there, but then again, when the hell is that useful?

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 4:11:36 PM

I'm trying to dl the vid, but why is their server so slow!? AAAAH!

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 5:24:28 PM

Yep, it doesn't beat the Ti4200. In same cases, not even close. Maybe 1 or 2 instances, but that's it. Toodles, point proven.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 11:11:30 PM

I rest my case with that review, when ati wins is wins by a hair, when it loses(most of the time) it loses by much more.


The 8500@275/275 with 128 megs of ram and catylyst drivers

DOES NOT BEAT THE TI4200!



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 19, 2002 11:22:55 PM

Quote:

The 8500@275/275 with 128 megs of ram and catylyst drivers

DOES NOT BEAT THE TI4200!


It matches the Ti4200 for less and has more features. You can get the R8500 128MB retail at NewEgg for $159!

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 19, 2002 11:25:52 PM

That's the thing, it DOESN'T match it, and to be honest, I could care less about tv in.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 19, 2002 11:29:04 PM

Looks like you have a untweaked broadband connection my 5.5mb connection got the video down fast and imagine I live on a caribbean island haha.
June 19, 2002 11:30:49 PM

Yeah, but it's cheaper, so for a person who can't shell out close to $200 for a GF4 Ti4200 128MB, the R8500 128MB could be quite useful. Especially with that 3.3ns RAM which I think everyone forgot to mention.

What's the deal with lampshades, I mean it's a lamp, why would you want a shade? :smile:
June 19, 2002 11:35:50 PM

You are very strange it beats the ti 4200 at 1600*1200 by a whooping 27% margin in jedi knight 2 and you call that winning by a hair and in some other benchmarks by 15 - 20%. Damn you are very demanding individual. By the way 3d mark 2001 benchmarks don't count. YOUR STATEMENT IS VERY PERPLEXING ACTUALLY I AM IN SHOCK REALITY HAS SUDDENLY BECOME VERY DISTORTED.
June 19, 2002 11:37:24 PM

Quote:

You are very strange it beats the ti 4200 at 1600*1200 by a whooping 27% margin in jedi knight 2 and you call that winning by a hair and in some other benchmarks by 15 - 20%. Damn you are very demanding individual. By the way 3d mark 2001 benchmarks don't count. YOUR STATEMENT IS VERY PERPLEXING ACTUALLY I AM IN SHOCK REALITY HAS SUDDENLY BECOME VERY DISTORTED.

THANK YOU! :smile:

:wink: <b><i>"A penny saved is a penny earned!"</i></b> :wink:
June 20, 2002 12:02:33 AM

Funny thing prices:
From newegg:
AIW R8500 128MB- $305
Ti4200 128MB- $172
Hmm

Now to the benchmarks (btw, the way I figure differential is R8500 catalyst/Ti4200 performance, and then that -1)
(negative indicates ti4200 better performance)
1600x1200 1280x1024 1024x768
3D mark 0.030645437 0.024701195 -0.005270926
Q3 -0.045172719 -0.047648515 -0.012652508
SS Extreme 0.201257862 0.174946004 0.082058414
SS Max -0.062642369 -0.073684211 -0.084848485
JK2 Max 0.270516717 0.14494382 0.041836735
Commanche 4 -0.101408451 -0.122596154 -0.120454545
Code Creatures -0.403225806 -0.38559322 -0.372093023
Look at the last one- realize that newer games coming out are geared towards the Gf4 series, and we'll be seeing more of that.

My frog asked me for a straw...dunno what happened his ass all over the place :eek: 
June 20, 2002 12:10:26 AM

Actually here is a quick break down of some of the benchmarks since your math skills are clearly lacking, Quake 3 at 1024*1280 ti 4200 wins by A VERY LARGE INCREDIBLE MARGIN OF 2.8% and at 1600*1200 by a colossal margin of 5.1%.

In serious sam extreme quality the radeon wins the benchamrks at 1024*1280 by a hair only by one hair(only by a sucky 20% margin).

In serious sam 2 at maximum quality (which is actually lower quality than extreme quality) at 1024*1280 the ti 4200 wins by AN UNBELIVABLE MARGIN OF 9.2% I am breathless help me someone. In commache 4 the ti 4200 wins by margins of 13.6% and 11.2% but in my opinion both cards render this game to slow actually not even the ti 4600 produces consist 60 fps. Oh well guess we have to wait for the rv250 to show how fast a game really should be (75% fact 25% speculation BITE ME FOR THIS STATMENT). Codecreatures does not count why because it is not a game it runs very slow even on the ti 4600 and is written specfically for the geforce 4. It is possible to produce a even prettier 3d demo running on r8500 hardware at a faster rate than on the ti 4600 by writting the demo from ground up on the r8500. IN other words codecreatures is just nVIDIA PR FLUFF.


One more thing matisaro did you see me say the 8500 beats the ti 4200.
June 20, 2002 12:14:42 AM

You silly ----- so what I could also find a website the sells the r8500 for less than the ti 4200 and why on earth are comparing the r8500 to the ti 4200 when the r8500 gives similar performance and can be had for more than $100 less YOU ARE DEFINTINELY A NVIDIOT.
!