Kneel and Repent; Jesus hates Nvidia! (And so do I)

Subtitle; 3D Mark is not the be all and end all when evaluating preformance.

I finally found the TNT that came with the computer I bought from Flam; seeing how it had 32mb of RAM, and the Voodoo III had 16, I thought that installing it could solve the 'white textures' problem; that it did, and I was pleased. Delta Force - Black Hawk Down, 3D Mark 2001 and Neverwinter Nights all functioned without the whiteness. However, out of idle curiousity, I ran 3D Mark 2000 on default benchmark. 1270 marks; I was pleased, normally with the Voodoo III I get 1250. Then I realised that I was using drivers that came out early '99; hmm... more preformance could be achieved, I thought to myself! A quick trip to the Nvidia website produced... updated drivers! Which I installed. Again, I ran 3D Mark 2000. 4522 marks. I was astounded. Jeebus Crust! What a fool I have been to support 3DFX even though they died years back, and then been bought by my arch-nemesis, Nvidia! For a second, my faith wavered.

Then, I thought, well, I might as well give it one last test, one last nail in the coffin of 3DFX fandom. I booted up CS, changed from 3DFX minidriver to default, *bumped* down the res from ~1200 to 1024, gave it a go. I typically get 40-80 frames per second on the Voodoo III running ~1200 x somebullcrap, so I was expecting four times that with the Nvidia card, after all, it had gotten almost that much in 3D Mark, correct? And, the res was lower.

My faith was restored. The Nvidia card achieved 14 frames a second, never exceeding 20, not even when I was looking at my own feet. I was flabbergasted; how could this happen? I checked the drivers, re-installed them twice, made sure I was in the correct GL settings, even gave it a shot in Direct 3D; all to no avail.

Conclusion; Nvidia is the lovechild of John Howard and the world's biggest vaccum cleaner; it can suck and blow at the same time, while telling you everything is perfectly ok.

Disclaimer: The author of this post does not claim any affiliation, influence, connection, disconnection, interpretation, relationship, attachment, membership, link, authority, sway, nor have bearing on, understanding, elucidation nor any other form of association with Jesus or the views or opinions of Jesus or any other holy deity, god, goddess, idol, figurehead, religions icon, totem, spirit, super, sub or un-natural being, nor the representation, incarnation, emblem, totem, model or religious concept.

On a serious note, does anyone know what could be wrong? Does Nvidia have the equivalent to a 'MiniGL driver' or something that optimizes it for OpenGL?

Quote from Wolfspirit:
.....I ran 3D Mark 2000. 4522 marks......

A TNT 2 out-performs a Voodoo5 in 3Dmark as well. Strange dontcha think? :lol:

I havent had any benchmarking programs except for sandra (only because someone wanted to measure the size of their penis against the speed of my hard drive) on my computer for ages and i dont want any on it.

Just goes to prove that Nvidia suck and blow. Their drivers are optimised to provide good scores in 3D Mark. Also with Nvidia you get a large amount of frames in most games (depending on the card) but you lose that in image quality, AA performance (in games) compared to 3dfx and probably ATI, and also lets not forget that Nvidia cards run jumpy in CS at anything below 50-60 frames, whereas a 3dfx card can go down to 30, and in the case of the VooDoo 5 25 frames before it starts to get jumpy.

Hey Wolfspirit, remember that every Voodoo2 or 3 card runs software AA. Disabling it may also incrase your performance by taking a load off your CPU, but if you have a decent CPU say 1GHz or near that then don't worry about it.
I remember what Locutus and I had TNT2s and we got the latest detonator drivers everybody was tossing over. The drivers were suposed to add 50% performance. Our 3DMark scores increased, but CS had dropped to around half the frames we used to get (reduction of 50%). :shock: But many people stuck to them just because they thought they were better because of 3DMark :shock: :shock:
My TNT2 outperforms my Voodoo5 in 3DMark :shock: , but the TNT2 can't hold a candle to the Voodoo5 in real games , the difference is unbelievable so don't base ANYTHINGon 3DMark scores whatsoever!

...all i have to ask is....

how the f*ck do you get 80fps in cs with a v3.

are we talking...online...proper online..like with players...?

shit i get 30-40 with my radeon...

(did get 100 until cs1.5....although i've been thinking recently that i did update my drivers at this point in time as well....:|)

WickedGL probably :P

Quote from Seqo_ae:
...all i have to ask is....

how the f*ck do you get 80fps in cs with a v3.

are we talking...online...proper online..like with players...?

shit i get 30-40 with my radeon...

(did get 100 until cs1.5....although i've been thinking recently that i did update my drivers at this point in time as well....:|)

Yeah same with me i was getting 100 fps with 2xAA till 1.5 then i had to turn off AA and i still only get 50-70. I think they fucked 1.5 up a bit, but then again they might have an agreement with hardware manufacturers to make it reuire more resources

1.5 is evil, I can still get 99fps with 2xAA, lod -0.75 and @1024*768*22 but it will drop to around 80 on some maps.

99fps isn't hard for a Voodoo3, you just need a decent CPU behind it. I was getting around 90fps in Elite Force (Q3) at 1024*768*32 (not using 3dfx GL enhancments)! Cs is more hardcore than Q3 these days thanks to some dodgy maps and mod stacking (making it inefficient).

I have a 533mhz, 320mb ram, Voodoo III. I get 80fps with no-one in it, and about 50-60 with seven bots (Realbot AI)...

CA drivers man. Rock0r

I'd would constantly get 99fps steady with my Voodoo 3 running 'the old wickedGL' drivers. Also running at 1152 x 864 resolution, P3-500, 500M+ RAM. It would go fine with a full 20 player server with 150ms ping, and everyone running around normally. However 2 smoke grenades in the air and lots of movement and shooting took stablility to new lows.

My V3 and V5 don't suffer from smokies so much, only with hardware or software AA (no shit, think about what it's doing) (quiet you).

I used to pull a fairly consistant 99-100fps (yes, it does read 100 in net_graph) on a generic gf2 MX400 with celeron 800, and to top it off, running under WinXP at 1152x864
The only time it would drop was when some1 used a smoke gren, dropped to about 40.
And all the time using the detonator drivers
Just make sure u use the older drivers. The new ones suck

Yeah my TNTPoo CD drivers were the best ones I ever had.

i am not a fanboy, but i wish to make you ppl aware of another point.

Just because its super fast in 3dmark, and its slow in CS doesnt mean its shit.

and just because its fast in CS doesnt mean its great.

And vice versa applies.

In this thread, all you have really discussed is CS, and 3dmark.

there is no comparison, primarly, 3dmark is a d3d benchmark, and CS is an opengl engine.
Another factor is that the engines in the different programs have been written with different apis in mind. Look at unreal.

Unreal will always look good, smashing even, on a 3dfx based card. Its the way of life. my old voodoo2 used to SCARE me with the graphic quality in unreal. Getting a tnt2 ultra pissed me totally off, because i lost it, i lost my great quality.

the unreal engine was written with GLide. that is all it is.

The engine that CS uses is a revamped q1 engine. It was originally a software only design, but it was ported to Glide (the card of that era) easily, and its always been a GLide engine. Thus, it will run very damn well on 3dfx cards.

I want you all with the 3dfx cards to tell me how well they perform on a d3d game. I dont know any of my head, but i am sure there are games out there that are written on a d3d core. And then tell me which is faster, a nvidia card or a 3dfx card.

nvidia has always been a d3d supportor. There cards have always done d3d very well. With engines that are not written for it, they suck.

fact of life.

True true. My Voodoo III preforms very badly in Warcraft III, even on 640x480 and low everything. But the simple fact is, I only use it to play Direct 3D games, so having great D3 support isn't really relevant to me. But, thanks for bringing in some moderation into that. :)

Depending on resolution the V4 and V5 somtimes run faster using D3D in Unreal Tournament! Glide just looks better. My V3 runs faster than my TNT2 in other games aswell, Elite Force, Star Trek Armada, Diablo II :shock: , and various others.

As for CS I really can't notice the difference between D3D and Glide, the V5 looks good in both it's just AA runs faster in Glide or OpenGL than in D3D, but seeing the difference in OpenGL and D3D on 'other cards' is so massive it's scary :?

nvidia AA is simple crap and you cant compare it.

go compare it with an ATI AA....

and the v4/v5 are faster then tnt2s. you would hope it did that.

the crap scores those cards get in 3dmark are probably attributed to shitty drivers on 3dfx's part, and not some hidden top secret scam by nvidia.

point to the company that has been updating its drivers nearly every month for the last 4 years?

Quote from rainwulf:
the crap scores those cards get in 3dmark are probably attributed to shitty drivers on 3dfx's part
They didn't care about 3DMark, but ATi did. Who wants to optimize for a non-game. Is 3DMark multiplayer?

Quote from Odyssey:
[quote:7feed32aaf="rainwulf"]the crap scores those cards get in 3dmark are probably attributed to shitty drivers on 3dfx's part
They didn't care about 3DMark, but ATi did. Who wants to optimize for a non-game. Is 3DMark multiplayer?[/quote:7feed32aaf]

i didnt mean drivers optimized for 3dmark, and you know it.

I meant the drivers being optimized for d3d at all. They had glide, and glide works like a charm, i dont think they cared too much about its performance in d3d.

and re-iterating my previous point, nvidias drivers at that time were probably the same. 3dfx havnt had the benefit of a dedicated section of a company devoted to drivers. you cant help things improving with time. 3dfx never had that time to optimized its drivers.

And i also found a good 3d game......serious sam.

benchmark the wickedgl, and then d3d in serious sam with your voodoos.

its a shame, but glide was a fucking incredible api. The only problem with it is that 3dfx never let anyone else licence it. they did an apple.

you are comparing two companies that are here today.. with one that isnt.
and you are complaining that the two modern companies "optimize" for 3dmark? did it ever occur to you that just MAYBE that because they are still around they have been updating their drivers (especially nvidia) for a lot longer then 3dfx?
and a side effect is *GASP* that you get "BETTER" performance with 3dmark?

what did you expect? it to get worse?

Quote from Odyssey:
[quote:5ee085679f="rainwulf"]the crap scores those cards get in 3dmark are probably attributed to shitty drivers on 3dfx's part
They didn't care about 3DMark, but ATi did. Who wants to optimize for a non-game. Is 3DMark multiplayer?[/quote:5ee085679f]

There fucking idiots then. Because like it or not people like to benchmark there systems vs. others.

( i dont really care about 3dmark much either )

Pity people are too interested in dick measuring contests rather than how well a component performs in an application that its used for.

'sif 3Dmark is a good benchmarking tool.

A good tool would provide a way to get exact statistics that let you know where the limitations are, providing you with a tool to test your tweaks and see exactly where the bottlenecks you need to address are.

3DMark provides some of that functionality, but I'd like to be able to get exact specifications on what my card is doing (as opposed to theoretical capabilities).

Give me the actual data transfer along the AGP bus, actual memory bandwidth, give me time to render in ms a shape of certain complexity with or without textures, lighting, environment mapping, etc. Show me what increasing that complexity does. No "All-in-one" benchmark score, real information, useful information.

Well the best benchmarks are the actual applications you use the hardware for innit :P

Quote from ruin:
Well the best benchmarks are the actual applications you use the hardware for innit :P

They give no help for performance tuning your setup or identifying performance issues.If you can identify problems you can work out what you can do to fix them.

They are the purpose you are tuning the hardware for though.

Benchmarks are for tweaking your system and for product comparisons. A benchmark like I mentioned earlier would give far more information than 3dmark.

But you can still check out exactly how much better your system runs where it counts when you tweak something.

Yes, but you can't easily identify the bottlenecks and therefore target your tweaking to fix them.

Interesting points.
I still believe that if people want to benchmark their hardware, they should do so in the programs/games they want to run. 3DMark isn't a game and it's performance figures shouldn't represent any D3D game.

Everybody know 3dfx card SUCK HARD when it comes to d3d. nuff said.
I'll bet that tnt2 will shit on the v3 in any d3d game.

Actually I find thay my V5 usually runs better in either Glide (no surprise) or D3D rather than OpenGL. Take UT for example. I don't know where people get that idea from, probably because Glide is similar to OpenGL.

as for glide, i am not sure what it actually is closer too...

doesnt matter though. its history now.

i prefer OPENgl for all my games anyway, they always run better in opengl.

D3d... in ss2, its.. laggy, move the mouse, and the image takes a few ms to follow... Beats me.

as for 3dmark not being a game.... well.... its actually using a game engine.
you can get a score just by doing the first 4 game tests.
and especially the first test, which actually IS a game demo ( you can play it), i dont know why ppl think that 3dmark is game related.

however it has been proven that both ati and nvidia fuck with benchmarks... its as simple as that.

it just comes down to what goes best. if your voodoo5 goes well in CS, then just use the fucking thing and stop telling others to change, or that your card is better.

I am quite happy that i can run CS at 75fps constant at 1280x1024, and that is with a gf3.

you cant blame companies that are still alive today for optimizing their drivers. its just a fact of life. You gotta imagine what 3dfx drivers would be like today, if they were still around. And soon od, the nv30 with its 3dfx parts will be out and maybe you will have to change your opinion.

as for the start of this thread, a person who decided that his tnt2 WAS faster, but then ran cs and saw that it sucked, did you consider all the other factors in this assesment? colour depth? ansitropic? AA even? LOD depth? what depth the textures were at? did you check OTHER games before making your assesment that nvidia blows goats?

you ran your voodoo and a tnt on an engine that was primarily written in software, and then ported to glide... there is a bias there.

I am sure a voodoo2 would kick a tnt2 in a game like Incoming.

i would like to see a series of benchmarks done in opengl and d3d, and in the case glide on a lot of those older games, and benchmarks, between a voodoo3 and a tnt2.

i for one used to HAVE a voodoo3, but i sold it and got a tnt2.

wanna know why?
because of 32 bit colour. banding sucks.

and my first assesment of this? damn. NFS3 doesnt look good anymore.
and Unreal doesnt look good anymore. damn all the games written for glide dont look as good anymore.

today, games are opengl or d3d. you dont have legacy glide games around anymore apart from good ol CS.

just discussion points. I for one hate bias and as soon as i sell my motorbike i am getting a 333 bus xp2800+, and an ATI 9700pro.

i am a fanboy of performance. no more 3dfx/nvidia wars please. its getting boring.

Quote from rainwulf:
...ansitropic?....LOD depth?..

On a TNT2, thats umpossible

Quote from rainwulf:
1. as for glide, i am not sure what it actually is closer too...
2. I am quite happy that i can run CS at 75fps constant at 1280x1024, and that is with a gf3.
3. And soon od, the nv30 with its 3dfx parts will be out and maybe you will have to change your opinion.
4. as for the start of this thread, a person who decided that his tnt2 WAS faster, but then ran cs and saw that it sucked.
5. did you consider all the other factors in this assesment? colour depth? ansitropic? AA even? LOD depth? what depth the textures were at?
6. i would like to see a series of benchmarks done in opengl and d3d, and in the case glide on a lot of those older games, and benchmarks, between a voodoo3 and a tnt2.
7. wanna know why? because of 32 bit colour. banding sucks. (meaning that na V3 doesn't have 32bit coulour)
8. damn all the games written for glide dont look as good anymore.

1. It's closer to OpenGL, however actually OpenGL is closer to Glide :wink:

2. OMG that's lame, an the GF3 is suposed to be how much faster than a V5? :oops:

3. I already plan on getting an NV30 (that must be the reason I have a computer rendered image created by an NV30 as my tag in CS :idea: ) because, yes it has 3dfx technology because nVidia got 100 3dfx enginers 8) and the NV30 is the first nVidia chip since the Riva 128! :shock: (not just a modified one) I don't really hate nVidia, only the TNT through to GF4 chips and some of the nForce northbridges.
BTW go and read the topic "NV30 Puzzle Coming Together - Slowly" by me. I think I have made it clear that I like the NV30 more than the R300.

4. "was", I don't think so, he said that 3DMark was higher, not a game so, no, you are wrong the TNT was slower.

5. AF and LOD on a TNT2 are you crazy? They had a form of AA however, I'm not sure if it was hardware or software. All I know is that it was poor image quality with a massive performance hit (just like the GF2).

6. V3 and TNT2 forget that, I have seen benchmarks between a GF2MX and two V2's in SLi with an XP1800. To put it simply the V2's beat the GF2MX in games designed around the time of the MX. People forget the scalability of 3dfx cards with faster CPUs.

7. Neither the TNT2 or V3 did true 32bit coulour :oops: (but the V3 did true 24bit tho). The GF256 and VSA100 were the first chips to do so. The TNT2 and V3 could both do artificial 32bit.

8. Maybe not on your old V3 card, remember the VSA100 was very much about image quality and a much newer chip.

I am running a Geforce2mx and was running a TNTpoo both were a complete waste of fucking money it's shithouse I can get better image quality running a Voodoo 3 for fuck sake why is it that nvidia is so crap can't they make a decent graphics card for fuck sake that legitimately get's decent frames and actually looks good :evil:

Quote from Lt.Cmdr. Tuvok:
I am running a Geforce2mx and was running a TNTpoo both were a complete waste of fucking money it's shithouse I can get better image quality running a Voodoo 3 for fuck sake why is it that nvidia is so crap can't they make a decent graphics card for fuck sake that legitimately get's decent frames and actually looks good :evil:

Whinge, whinge, bloody whinge. Sorry, but I get pissed off when people crap on about how a _ancient_ card is so called "crap". What were you expecting? Your TNT to be throwing out Toy Story 2 rendered on the fly?

And the GF2 MX is a budget card. I had a GF2 MX for ages, damn good card for what it does. Pretty cheap, does the job. Image quality is ok, but hey, when I'm running around in a game fragging people, I don't sit there going "Oooh, aahhh, see how well anti-aliased the edge of that wall is? And I only had to drop 20fps to get it so lovely and smooth!"

And nVidia CAN make a decent graphics card that gets decent frames and looks good. ANY accelerated card is going to produce reasonable picture quality. Hell, my Banshee made an ok picture :) If you want to remember what a pixel storm is, go back to software mode :)

I agree with Martok and Odyssey that the current nVidia range is nothing more than seriously overclocked TNT's with a few extra features tacked on. So what? Your fancy Athlon is still emulating the same basic setup that the ol' 386 used, which was just a bunch of extensions on the ancient 8086 :) However, I am looking forward to when nVidia finally coughs up a GPU that's been totally redesigned, ie, the NV30.

I'm not an nVidia fanboy, nor am I a 3dfx diehard, nor an ATi whore. Show me a good card, and I'll consider it. I just get sick of people bitching at nVidia all the time.

Bloody tall poppy syndrome again.

Darkwolf.

Quote from Darkwolf:
And the GF2 MX is a budget card. I had a GF2 MX for ages, damn good card for what it does. Pretty cheap, does the job. Image quality is ok, but hey, when I'm running around in a game fragging people, I don't sit there going "Oooh, aahhh, see how well anti-aliased the edge of that wall is? And I only had to drop 20fps to get it so lovely and smooth!"

I agree with Martok and Odyssey that the current nVidia range is nothing more than seriously overclocked TNT's with a few extra features tacked on. So what? Your fancy Athlon is still emulating the same basic setup that the ol' 386 used, which was just a bunch of extensions on the ancient 8086 :) However, I am looking forward to when nVidia finally coughs up a GPU that's been totally redesigned, ie, the NV30.

Firstly, games suck if they look shit, secondly AA on a GeForce 2 should be called vas vision. Also on the CPU part, its totally different because a CPU doesnt render images, it crunches numbers, therefore the higher the clock the better it will be. 3d chipsets dont have nearly as much benefit by pure overclocking. (unless its an Nvidia card tested in 3dmark)

Quote from Darkwolf:
when I'm running around in a game fragging people, I don't sit there going "Oooh, aahhh, see how well anti-aliased the edge of that wall is? And I only had to drop 20fps to get it so lovely and smooth!"
You loose 20fps when using AA?!?
How can you miss the image quality when the entire screen is flickering and shimmering? The only way a GF2 (any) can look smooth is if you enable their VasVision AA or use the default LOD that nVidia provides :shock:
Anyone can blurr the image as much as they do.

Quote from Odyssey:
You loose 20fps when using AA?!?

I have no idea how many FPS I lose when I turn on AA. It was an example. Only ever turned it on a couple of times in a game, didn't notice the difference, so turned it back off again.

Examining some picture quality snaps from 3dmark, yeah, I can see the difference between AA and non-AA, but seriously, if I'm racing about trying to frag some bugger, the last thing I'm looking for is jaggies. And I had to zoom in on those snapshots in Photoshop to get a good example of the difference.

Darkwolf.

He probably has vsync turned on. hence 75 max refresh his monitor can support..

Quote from Darkwolf:
Only ever turned it on a couple of times in a game, didn't notice the difference, so turned it back off again.

Huh? You didnt notice Vas Vision? :P

Quote from Darkwolf:
Examining some picture quality snaps from 3dmark, yeah, I can see the difference between AA and non-AA.
A still picture won't really show the difference, yeah there will be no jagged edges, but the shimmering/jittering while moving is the main reason for using AA in games like CS (and don't tell me the nVidia cards don't do this because I know better). This can be also solved (slightly) by using a default (zero) or positive LOD, but who wants that? (obviously you guys).

Quote from Flammable:
He probably has vsync turned on. hence 75 max refresh his monitor can support..
Are you talking about me? If so my monitor can do 87Hz max.

Quote from MARTOK:
Huh? You didnt notice Vas Vision? :P

Wtf is that?

Anyway, here's a snapshot of CS with and without AA...

Specs: de_piranesi, 1024x768x32, vsync on, ASUS V8460 Geforce4 Ti4600 Ultra Deluxe, using OpenGL, nVidia Detonator 40.71 Beta.
On the left: 4x AA, 8x Anisotropic filtering.
On the right: No AA, no anisotropic filtering.

Summary:

Again, I don't notice any "shimmering" effect in CS, or any other game for that matter. But, for some reason, my bullets seemed to be skull magnets with the 4x AA (got 74 headshots in 100 kills). Performance hit is quite low in CS - but CS is easy on the GPU. Will try later in BF1942 and measure the performance hit.

Darkwolf.

Quote from Odyssey:
[quote:0d812effe6="rainwulf"]1. as for glide, i am not sure what it actually is closer too...
2. I am quite happy that i can run CS at 75fps constant at 1280x1024, and that is with a gf3.
3. And soon od, the nv30 with its 3dfx parts will be out and maybe you will have to change your opinion.
4. as for the start of this thread, a person who decided that his tnt2 WAS faster, but then ran cs and saw that it sucked.
5. did you consider all the other factors in this assesment? colour depth? ansitropic? AA even? LOD depth? what depth the textures were at?
6. i would like to see a series of benchmarks done in opengl and d3d, and in the case glide on a lot of those older games, and benchmarks, between a voodoo3 and a tnt2.
7. wanna know why? because of 32 bit colour. banding sucks. (meaning that na V3 doesn't have 32bit coulour)
8. damn all the games written for glide dont look as good anymore.

1. It's closer to OpenGL, however actually OpenGL is closer to Glide :wink:

2. OMG that's lame, an the GF3 is suposed to be how much faster than a V5? :oops:

3. I already plan on getting an NV30 (that must be the reason I have a computer rendered image created by an NV30 as my tag in CS :idea: ) because, yes it has 3dfx technology because nVidia got 100 3dfx enginers 8) and the NV30 is the first nVidia chip since the Riva 128! :shock: (not just a modified one) I don't really hate nVidia, only the TNT through to GF4 chips and some of the nForce northbridges.
BTW go and read the topic "NV30 Puzzle Coming Together - Slowly" by me. I think I have made it clear that I like the NV30 more than the R300.

4. "was", I don't think so, he said that 3DMark was higher, not a game so, no, you are wrong the TNT was slower.

5. AF and LOD on a TNT2 are you crazy? They had a form of AA however, I'm not sure if it was hardware or software. All I know is that it was poor image quality with a massive performance hit (just like the GF2).

6. V3 and TNT2 forget that, I have seen benchmarks between a GF2MX and two V2's in SLi with an XP1800. To put it simply the V2's beat the GF2MX in games designed around the time of the MX. People forget the scalability of 3dfx cards with faster CPUs.

7. Neither the TNT2 or V3 did true 32bit coulour :oops: (but the V3 did true 24bit tho). The GF256 and VSA100 were the first chips to do so. The TNT2 and V3 could both do artificial 32bit.

8. Maybe not on your old V3 card, remember the VSA100 was very

much about image quality and a much newer chip.[/quote:0d812effe6]

mm how do i go about replying? hahaha
ok lets see.
point 1
yea i was pretty sure opengl and glide are close, but there are quite a few differences, thats why you had to get miniports for opengl games to work on voodoo cards.

point 2.
a gf3 is a shit load faster. i said a constant 75 fps. so what? its called vsync dumbass. i lock CS at 75? do you have a problem with that? its capable of doing a shit load more, but 75 does me just fine, and CS has a inbiult limit of 99 fps.
so dont go getting all snotty nosed about how superiour your voodoo is, if it was so superior, your beloved company would still be alive eh.

point 3
shrug. i buy whats fastest. as for the geforce series, the tnt and geforce2 are different architectures, however from the gf1 up, its all basically the same. we all knew that. its not a point you have to yell at us for a proof that nvidia is evil. its called ramping. all companies do it. get something good, make it go faster. Its too expensive to redesign architecture every 6 months you idiot. the voodoo series is exactly the same so shut up.

point 4. i may as well just repeat my point. idiot. he decided his tnt was faster due to 3dmark. its there in the thread. dont misquote me and then claim a victory. its a lack of intelligence.

point 5. all nvidia cards have LOD. its just a texture filter.

point 6. i got a tnt2 ultra cos it was faster then a voodoo3. and it also looked better. sorry. go look up the q3 benchmarks. and 32 bit wins all the time.
the voodoo3's FAKE 22bit colour was shit. and YES thats all it was capable of. 22 bit internal rendering, and ONLY ever 16 bit output. I dont think anyone anywhere LIKES 16 bit rendering because it was ugly as hell.

point 7. the TNT2 DOES do true 32 bit colour. it was one of its selling points. dumbass. even the TNT did 32 bit rendering.
here are the specs on an M64..
http://www.siig.com/products/video/specs/AuroraAGPTNT2M64_S.html
the voodoo3 never had anything to do with 32bit anywhere. ever. it didnt even support agp properly.

point 8. i didnt buy a vsa card though did i. The nvidia series looked so much better. The voodoo3 series was out, and the tnt2 ultra was beating it in all games apart from natural glide games. And it looked better.

and to rub it in more another link to a tnt2 series product.
http://www.active-hardware.com/english/reviews/video/erazor3.htm

you can preach all you want. i dont know WHAT you are preaching though.. is it that 3dfx is better? that there cards scale better with CPU?
doesnt that kinda of indicate that there were pretty poorly designed to start with?

i would really love it if CS wasnt the only game used for comparison here. We all know that 1.5 has got problems, and that the type and age of the engine plays major havock with results.

you play CS at what? 1280x1024? i dont know, tell me.
and you say its so smooth at 25fps......... um... 25 is slow. can it do 1280x1024 at 100 fps?
i dont even know if a gf3 can do that, will check now.

ok cant make CS ignore the vsync dammit.
well jsut then it was running at 1600x1200 at 60fps constant. that must mean i have a shit card. :?

oh.. fixed it.

1600x1200 32bit colour.... and 100/99 fps constant.
who NEEDS AA? lol

and if anyone wants proof, i got the screeny. i tell you what, at that res everything is so clear....... and smoooooooooooth.

mm got bored.

1600x1200 4xAA and 16 bots having goes at each other........
70 fps constant. mmmmmmmmmmm

gf3 ti500 is more powerful then i thought.

Quote from rainwulf:
a gf3 is a shit load faster. i said a constant 75 fps. so what? its called vsync dumbass. i lock CS at 75? do you have a problem with that? its capable of doing a shit load more, but 75 does me just fine, and CS has a inbiult limit of 99 fps.

.....its called ramping. all companies do it. get something good, make it go faster. Its too expensive to redesign architecture every 6 months you idiot. the voodoo series is exactly the same so shut up.

the voodoo3's FAKE 22bit colour was shit. and YES thats all it was capable of. 22 bit internal rendering, and ONLY ever 16 bit output. I dont think anyone anywhere LIKES 16 bit rendering because it was ugly as hell.....the TNT2 DOES do true 32 bit colour. it was one of its selling points. dumbass. even the TNT did 32 bit rendering.

...voodoo3 didnt even support agp properly.

...there cards scale better with CPU?
doesnt that kinda of indicate that there were pretty poorly designed to start with?

Firstly if your an nVidia boy that doesn't care if vsync is limiting your frames then you obviously don't have penis envy like some other people I bag *looks at somebody in these forums* and that's a good thing because your not as sad as them.

Ramping you say, why change an architecture that already has good performance and image quality. People said the TNT2 had poorer image quality than the V3 way back and its not until now that nVidia has bothered to ask external people about what they think of nVidia quality, another thing the NV30 will be better than any existing nVidia product.

32bit true colour, I don't think so, nVidia said their first chip to do true colour is the GeForce256. 16bit is crap, in OpenGL and especially DX, but not Glide. I never run it myself as I'm either using 22 or 32, except glide, but that may also do 22 with AA I'm not sure.

No a V3 didn't support AGP at all. Nor does my V5. Its simply uses the AGP as if it were a 66MHz PCI slot. I'm not entirely sure, but I think the V4 did use AGP properly as it could do 4x when the V5 5500 did 2x (I've been told the 6000 can do 4x 3.3v though).

You said that these cards were poorly designed if they scale so high with a decent CPU. That's not true, think about T&L. If it's done in software on the Voodoo series then wouldn't the performance jump with a CPU upgrade? The HW T&L on a GF2 was faster than software on a Pentium2, but with a Penium3 600 the V5 software T&L was faster. I don't know about you, but I don't use a P2 or even a P3 600 any more. The T&L with a P2 was superior, but now (because its performance is fixed) its now the GF and GF2 bottleneck.

My point is that in the time I have had my V5, people have been telling my that the "equivalent" nVidia products were better. Those people are still telling me that however they have upgraded not once, not twice, but three times in the time I have keep my V5. Doesn't that say something? The card is getting old as its three years old and you keep saying that I hate nVidia, but the truth is I like the NV30, just not anything between the Riva128 and GF4 and I don't like the kind of people that buy them, they are all the same (you enable vsync though so your not a typicial 'performance is everything' bogan).

If you think a GeForce2 or 3 (I'm talking to anyone) is so much faster than a V5 then you had better check this out.
BTW the V5 doesn't have any form of HSR (unlike the GF3) so it's overdraw is higher than a GF3 and the V5 is overclocked from 166 to 185 in some tests, not much but worth pointing out.

They were using the evil 1.04.01 drivers, that's why Quake3 ran so shit and the Thunderbird doen't have SSE. Running many programs especially ones with T&L the GF3 should beat it as software T&L on MMX or 3DNow isn't that great.

The V5 at 1600x1200 x2 FSAA 16bit seems to be the best option for the V5, what were you saying about 1600x1200 Rainwulf?

http://www.papadoc.net/V5vsGF3.html

Of the 39 benchmark numbers I have 18 were wins for the Geforce 3, 13 were wins for the Voodoo5 and there were 8 ties (tie being defined as anything within 2 frames), that's almost as astounding as the Nvidia demos. Who would have thought that a 3-year-old card would keep up that well with basically brand new card (at the clocked speeds the GF3 achieved its darn close to the Ti500). Just goes to show that you need to carefully analyze the games you play before buying hardware. In general you will be better off buying a faster CPU than replacing your V5, for instance going from a 750mhz P-III to a Athlon 1.4 almost doubled frame rates in Falcon, and was responsible for increasing on average my frame rates 40%!
The Voodoo 5's FSAA is simply in another class compared to the Geforce 3 and this is coming from a person who played the Geforce 3 for sometime before even starting to do this review to get a feel for it. I didn't think there was that much difference until I put two computers side by sideand ran the same missions on both cards at the same exact moments.

3D Labs is a leader in state of the art 3D Professional Accelerators. (note: High resolution does not cure aliasing as can be easily seen when comparing 1600x1200 images with and without FSAA, remember that screenshots only show edges)
Texture swimming and creepy crawlies, the worst offenders of aliasing, could NOT be shown by screenshots.

bah image quality. i never noticed any quality difference between my voodoo 5 and my geforce 3. infact any other geforce i have owned.. Maybe you cheap fags should spend some money on your monitor and you will see what your video card can put out..

BTW i run Vsync on. as my monitor can handle higher refresh rates than most monitors.

Quote from Flammable:
bah image quality.
BTW i run Vsync on. as my monitor can handle higher refresh rates than most monitors.
Your one of those 3d card bogans I was talking about.
Who cares about my monitor refresh rate. I am happy with it maxing my frames at 60. Why do you need to do this, does your card run jumpy at 60? Must do or otherwise you wouldn't change it, or maybe you have 'penis envy'.
30fps is also sweet. 8)

lol 3dfx and nVidia employees are still bitching;
When nVidia got ahold of 3dfx's technology from the buyout, they mixed nVidia and (former) 3dfx staff. They broke them up into two groups and had each group argue for different technologies to be included as a part of a radically new video card product (NV30). One group argued for technology nVidia had been culturing in-house while the other argued for technology created by 3dfx engineers. In this way, former 3dfx employees and longtime nVidia veterans collaborated with one another on creating the most revolutionary and arguably effective product possible.

Quote from Odyssey:
Your one of those 3d card bogans I was talking about.
Who cares about my monitor refresh rate. I am happy with it maxing my frames at 60.

Now, now, no need to be like that :) I don't know if it's a function of my card, or just my eyesight, but to me, 60 Hz looks like a strobelight. I can visibly see the flashing when on the desktop, and playing a game is like watching a slideshow (not that bad, but it looks really jittery to my eyes).

Hell, my eyes can even pick up the difference between 85Hz and 100Hz. 100Hz is nice and smooth.

That said, I _do_ run with vsync on - I can't stand the tearing effect that happens when vsync is turned off. Since my monitor can do 100Hz at 1024x768, that's what I play at.

Darkwolf.

Quote from Darkwolf:
But, for some reason, my bullets seemed to be skull magnets with the 4x AA (got 74 headshots in 100 kills
Isn't that a good thing? :wink:
Did you notice that 'flash' time was reduced with AA, or is that just a 3dfx thing?

Quote from Odyssey:
[quote:cb639d6225="Darkwolf"]But, for some reason, my bullets seemed to be skull magnets with the 4x AA (got 74 headshots in 100 kills
Isn't that a good thing? :wink:
Did you notice that 'flash' time was reduced with AA, or is that just a 3dfx thing?[/quote:cb639d6225]

Hehe, if I can get HS accuracy like that with 4x AA, then call me Captain Anti-Aliasing :)

Nah, I tried that stunt out - no change in flash time.

Darkwolf.

your a dick. it doesent look jumpy at 60. if your monitor can do higher than 60hz then why the fuck wouldnt u put it above 60. with higher frame rates you have an advantage in cs. with recoil disabpating quicker and smoother mouse. as well as a all round smoother feeling.

"Hell, my eyes can even pick up the difference between 85Hz and 100Hz. 100Hz is nice and smooth.

That said, I _do_ run with vsync on - I can't stand the tearing effect that happens when vsync is turned off. Since my monitor can do 100Hz at 1024x768, that's what I play at. "

I completely agree with that statement and concur whole hartedly. *spelling*

Once again your post means nothing to me (I do understand it, but don't care). You must be one of (if not THE) most penis envy person I have ever met.

You get an advantage when using low res because it's easier to HS people, but do you see everybody doing that? Hell no it liiks shit.

Why do I care about a tactical advantage as my most used wepons in CS are Mac10 and TMPs. If I really cared about winning I wouldn't use them.

I play CS for fun and eye candy. :wink:
[Image: http://www.pnl.gov/nsd/commercial/fmd/fire.jpg]

ohhh fuck it. your a dick and youll never change. i cant be fucked posting in this thread saying the same shit over and over again, just cause some fuck doesent understands and says i have penis envy..

Down boy :P
[Image: http://www.plainsboropolice.com/images/fire%20extinguisher.jpg]

Quote from Odyssey:
Down boy

:lol:

THIS IS RAINWULF

Quote from Odyssey:
[quote:43fa1a65c0="rainwulf"]a gf3 is a shit load faster. i said a constant 75 fps. so what? its called vsync dumbass. i lock CS at 75? do you have a problem with that? its capable of doing a shit load more, but 75 does me just fine, and CS has a inbiult limit of 99 fps.

.....its called ramping. all companies do it. get something good, make it go faster. Its too expensive to redesign architecture every 6 months you idiot. the voodoo series is exactly the same so shut up.

the voodoo3's FAKE 22bit colour was shit. and YES thats all it was capable of. 22 bit internal rendering, and ONLY ever 16 bit output. I dont think anyone anywhere LIKES 16 bit rendering because it was ugly as hell.....the TNT2 DOES do true 32 bit colour. it was one of its selling points. dumbass. even the TNT did 32 bit rendering.

...voodoo3 didnt even support agp properly.

...there cards scale better with CPU?
doesnt that kinda of indicate that there were pretty poorly designed to start with?

Firstly if your an nVidia boy that doesn't care if vsync is limiting your frames then you obviously don't have penis envy like some other people I bag *looks at somebody in these forums* and that's a good thing because your not as sad as them.

Ramping you say, why change an architecture that already has good performance and image quality. People said the TNT2 had poorer image quality than the V3 way back and its not until now that nVidia has bothered to ask external people about what they think of nVidia quality, another thing the NV30 will be better than any existing nVidia product.

32bit true colour, I don't think so, nVidia said their first chip to do true colour is the GeForce256. 16bit is crap, in OpenGL and especially DX, but not Glide. I never run it myself as I'm either using 22 or 32, except glide, but that may also do 22 with AA I'm not sure.

No a V3 didn't support AGP at all. Nor does my V5. Its simply uses the AGP as if it were a 66MHz PCI slot. I'm not entirely sure, but I think the V4 did use AGP properly as it could do 4x when the V5 5500 did 2x (I've been told the 6000 can do 4x 3.3v though).

You said that these cards were poorly designed if they scale so high with a decent CPU. That's not true, think about T&L. If it's done in software on the Voodoo series then wouldn't the performance jump with a CPU upgrade? The HW T&L on a GF2 was faster than software on a Pentium2, but with a Penium3 600 the V5 software T&L was faster. I don't know about you, but I don't use a P2 or even a P3 600 any more. The T&L with a P2 was superior, but now (because its performance is fixed) its now the GF and GF2 bottleneck.

My point is that in the time I have had my V5, people have been telling my that the "equivalent" nVidia products were better. Those people are still telling me that however they have upgraded not once, not twice, but three times in the time I have keep my V5. Doesn't that say something? The card is getting old as its three years old and you keep saying that I hate nVidia, but the truth is I like the NV30, just not anything between the Riva128 and GF4 and I don't like the kind of people that buy them, they are all the same (you enable vsync though so your not a typicial 'performance is everything' bogan).[/quote:43fa1a65c0]

thanks for the comment concerning vsync haha

as for the tnt2 vs v3 image quality, "People said the TNT2 had poorer image quality than the V3 way back " well other ppl said the opposite. Personally, i had both cards in my sticky hands for a few weeks, and in every game the tnt2 won out, no banding, no 16 bit rendering. from my own personal viewpoint, the v3 lost so much with image quality.
And i still stand by the fact that tnt2s did 32 bit rendering. I would like to see specs that say that tnt2 DIDNT do 32 bit rendering.

as for the t&l side of things, that is an unknown. i am not sure if t&l will always be used, i think you can turn it off.

and games, i dont think many games actually use T&L... so you cant use that as an argument, but i do say that i am prepared to be corrected on that.

it just comes down to use what you like. its simple.
fights and arguments wont change anyones opinoins, i am happy with Cs at 100fps at 1900x1200 ;)

end of thread.

Enough's enough

I didn't mean for this to turn into a flame war, which is what it has become. Thank you all for your help, but... yeah... taken a bit too far I think. Everyone needs to take a big step back and just say; it's just a video card.