What's new

VooDoo 5 5500 PCI info needed

fivefeet8

-= Clark Kent -X- =-
AlphaWolf said:


EDIT: Ok I just looked it up, the little sold Voodoo 5 6000, which was the best of the 3dfx pack ever made, barely matches that of a GF2 Ultra, which now costs peanuts. The V3-5 users would definitely benefit from a GF3 Ti200, which indexes at $64, and has all of your DX8/9 features.

Geforce3 cards aren't directx9 cards.. They are directx7/8 and lack the feature for directx9.. Only directx9 cards are geforcefx's and Radeon 9x00's..

They may be dx9 compatible but not compliant.. The voodoo 5 has superior FSAA quality compared to the geforce3 and especially the geforce2 in emulation.. Plus running with glide plugins, it's easily as fast and sometimes faster than OGL and d3d running on geforce2's and 3's in psx and n64 emulation..

The V5 6000 has higher memory bandwidth than the geforce2 ultra, but lacks certain features that the geforce2 has.. But it has 8xFSAA which would blow away the geforce2's crappy FSAA quality...
 

AlphaWolf

I prey, not pray.
fivefeet8 said:
Geforce3 cards aren't directx9 cards.. They are directx7/8 and lack the feature for directx9.. Only directx9 cards are geforcefx's and Radeon 9x00's..


Ok 8 sorry, but V5 isn't.


They may be dx9 compatible but not compliant.. The voodoo 5 has superior FSAA quality compared to the geforce3 and especially the geforce2 in emulation.. Plus running with glide plugins, it's easily as fast and sometimes faster than OGL and d3d running on geforce2's and 3's in psx and n64 emulation..

I don't know anything about the V5s FSAA, but I do know that the card is slower than a GF3, and having a card that slow, using FSAA isn't exactly a good idea.
 

milen

New member
The method used in 3dfx's FSAA is different than Nvidia's. Voodoo 5 6000 FSAA is much faster and better quality than Geofrce2/3.

I mean
Voodoo5 6000 slower than Geforce3 Ti200
Voodoo5 6000 4xFSAA faster than Geforce3 Ti200 4xFSAA
 

Reznor007

New member
AlphaWolf said:
That can't be right, the quincunx mode barely takes any performance hit at all on older stuff.

On VSA-100, the FSAA performance is easy to predict. 2xFSAA gives you 1/2 speed. 4xFSAA gives you 1/4 speed.

The only exception is the Voodo5 6000, which was 1/2 speed for 4x, and 1/4 for 8x. I'm not getting into the Quantum3D variants using up to 32 VSA-100 chips(and 2GB video RAM), which could do 64x FSAA at 1/4 normal speed.

The only card to really beat the VSA-100 FSAA is the Radeon 300 series(9700, 9500, etc.).
 

Tagrineth

Dragony thingy
Reznor007 said:
The only exception is the Voodo5 6000, which was 1/2 speed for 4x, and 1/4 for 8x. I'm not getting into the Quantum3D variants using up to 32 VSA-100 chips(and 2GB video RAM), which could do 64x FSAA at 1/4 normal speed.

Wrong.

4x would be 1/2 speed compared to a 5500's non-AA. :)
 

fivefeet8

-= Clark Kent -X- =-
AlphaWolf said:

I don't know anything about the V5s FSAA, but I do know that the card is slower than a GF3, and having a card that slow, using FSAA isn't exactly a good idea. [/B]

It's memory bandwidth is lower than a Geforce3, but in N64 and Psx emus, both the v5 and the geforce3 has more than enough of it to use FSAA and get good speed..

Hell, I can run psx and N64 games with glide at 960x720 + 4xFSAA on my V5 5500 PCI and stil get full speed.. And the FSAA quality is better than the Geforce3..

FYI.. The voodoo 5 5500 pci has the same memory bandwidth as a Geforce2 Ultra..

AlphaWolf said:
That can't be right, the quincunx mode barely takes any performance hit at all on older stuff.

Yeah.. Quincunx FSAA was designed for minimal performance loss, different form of FSAA than the other modes like 2x and 4x in geforce's, and looks like crap in anything 1024x768 and lower.. It only looks decent at higher resolutions..
 
Last edited:

Reznor007

New member
Tagrineth said:
Wrong.

4x would be 1/2 speed compared to a 5500's non-AA. :)

First off, I wasn't comparing 6000 speeds to 5500.

And we were both wrong. 4x on the 6000 would be 1/4 normal speed, as each chip renders it's own frame, thus having 4x as much to render.

Actual rates:
Voodoo4:
Normal=333Mpixel
2xFSAA=1/2 speed-166Mpixel

Voodoo5 5500:
Normal=666Mpixel
2xFSAA=1/2 speed-333Mpixel
4xFSAA=1/4 speed-166Mpixel

Voodoo5 6000:
Normal=1333
2xFSAA=doesn't work
4xFSAA=1/4 speed-333Mpixel
8xFSAA=1/8 speed-166Mpixel

4x on the 6000 is the same as 2x on the 5500, not full speed 5500.
 

Tagrineth

Dragony thingy
Reznor007 said:
And we were both wrong. 4x on the 6000 would be 1/4 normal speed, as each chip renders it's own frame, thus having 4x as much to render.

Which is exactly what I said :)

Actual rates:
Voodoo4:
Normal=333Mpixel
2xFSAA=1/2 speed-166Mpixel

Voodoo5 5500:
Normal=666Mpixel
2xFSAA=1/2 speed-333Mpixel
4xFSAA=1/4 speed-166Mpixel

Voodoo5 6000:
Normal=1333
2xFSAA=doesn't work
4xFSAA=1/4 speed-333Mpixel
8xFSAA=1/8 speed-166Mpixel

Nice job stating the obvious.

And ask any 6k owner about 2x AA. :) Every one of them will tell you 2x works, even if 4x and 8x crash. :)

4x on the 6000 is the same as 2x on the 5500, not full speed 5500.

Re-read what I said:

4x would be 1/2 speed compared to a 5500's non-AA.

4x on a 6k would be half as fast as a 5500's non-AA.
 

Reznor007

New member
Tagrineth said:
Which is exactly what I said :)



Nice job stating the obvious.

And ask any 6k owner about 2x AA. :) Every one of them will tell you 2x works, even if 4x and 8x crash. :)



Re-read what I said:

4x would be 1/2 speed compared to a 5500's non-AA.

4x on a 6k would be half as fast as a 5500's non-AA.

That isn't exactly what you said though. You only said the part about V5 6000 4x speed in relation to V5(which I wasn't comparing anyway). You didn't mention ANYTHING about how it works :)

Also, when you said that, I thought you were saying 4x on the 6k was the same as full speed V5 (half speed V5 6k).

And from everything I've heard 2x on the 6k didn't work due to either software bugs or flaws in the hardware(and let's face it...the 6k was a mess). I heard something about getting 2x to work, however that was in single chip mode, which would make a an expensive Voodoo4.

PS, that was only obvious if you knew about VSA-100 in the first place.
 

Tagrineth

Dragony thingy
Reznor007 said:
That isn't exactly what you said though. You only said the part about V5 6000 4x speed in relation to V5(which I wasn't comparing anyway). You didn't mention ANYTHING about how it works :)

Nitpicker. :flowers:

And from everything I've heard 2x on the 6k didn't work due to either software bugs or flaws in the hardware(and let's face it...the 6k was a mess). I heard something about getting 2x to work, however that was in single chip mode, which would make a an expensive Voodoo4.

A good number of 6k's can ONLY run 2x. The other modes result in a PCI bus desync and lockup within about ~5 minutes, without Hank's mod.

I haven't heard anything whatsoever about any 6k's being incapable of 2x AA. And I've been a part of the x3dfx community since the beginning. And I was a part of 3dhq from its start to its end. :)
 

AlphaWolf

I prey, not pray.
Now you all see why I don't like to pay attention to the technical side of 3D hardware. It's all heresay from sales asses and weirdos who measure their pride based on what video card they own. There isn't any accurate information anywhere.
 

Reznor007

New member
Tagrineth said:
Nitpicker. :flowers:



A good number of 6k's can ONLY run 2x. The other modes result in a PCI bus desync and lockup within about ~5 minutes, without Hank's mod.

I haven't heard anything whatsoever about any 6k's being incapable of 2x AA. And I've been a part of the x3dfx community since the beginning. And I was a part of 3dhq from its start to its end. :)

Well, the 2x works, but I've only heard of it working in single chip mode(otherwise it would be all 4 chips working together, which causes problems on most since they are all alpha and beta hardware).

I also follow the x3dfx stuff, since the day it opened, all the way up to the past day or 2 with the driver source released(I'm still wondering if that is completely true or what though, and yes, I have the file).

And I still feel the way I said about 3dhq...all hype.
 

Tagrineth

Dragony thingy
Reznor007 said:
Well, the 2x works, but I've only heard of it working in single chip mode(otherwise it would be all 4 chips working together, which causes problems on most since they are all alpha and beta hardware).

Talk to LogicalMadness or Colourless.

And all the performance figures I've seen show 2x running full speed. :flowers: I have no idea where you get the idea that 2x runs in single-chip only.

I suppose it might make sense if it ran on two chips only, considering that it doesn't fuck over the PCI signal like 4x/8x on most 6k's... but that doesn't make sense because all the numbers I've seen have it faster than a 5500 when both are in 2x.
 

Reznor007

New member
Tagrineth said:
Talk to LogicalMadness or Colourless.

And all the performance figures I've seen show 2x running full speed. :flowers: I have no idea where you get the idea that 2x runs in single-chip only.

I suppose it might make sense if it ran on two chips only, considering that it doesn't fuck over the PCI signal like 4x/8x on most 6k's... but that doesn't make sense because all the numbers I've seen have it faster than a 5500 when both are in 2x.

I suppose I could ask Colourless, but at this point it doesn't really matter.

But logically thinking, if 2x uses all 4 chips, it should suffer the same problems as 4/8x since all of the chips are functioning.
 

Tagrineth

Dragony thingy
Flash said:
I have 6K and have no problems with all modes, 2x, 4x, 8x.

Lucky! Most working non-final revision 6k's have severe problems with 4x/8x locking up within 5 minutes...

How's the performance in 2x compared to standard and 4x?

And, for the love of the Goddess, post some 8x screenshots!!
 

Top