Hi guys!

Surfin` some days ago, i have founded these thinghs:

3Dfx Confirms Voodoo2 Rev319:36 | by ChrisSteve Schick of 3Dfx has confirmed the rumors and speculation about the upcoming Voodoo2 Rev 3 boards, which are toted to have the capability to house more memory per board, and utilize 125 MHZ SDRAM instead of the current 100 MHZ bus SDRAM on current Voodoo2's. Thanks Agn for the news bit. 3Dfx has developed a revision to the Voodoo2 Reference Board which makes better utilization of memory and therefore runs faster. The chipset remains unchanged. The board does not have any new features, just the optimization of memory. The revised board can run over 100 MHz and can offer an overall performance boost of around 15 percent. So far none of the companies making Voodoo2 boards have adopted our new design revision. Steve Schick3Dfx Interactive, Inc.

And....

Another...

SLI with Different V2 Brands/Sizes09:21 | by ChrisAfter swapping out various Voodoo2 based cards from various systems, I decided to try a little experiment, to get 2 different Voodoo2 cards from different manufacturers to work together in SLI (Scan Line Interleave) mode. 3Dfx has said time and time again that you need identical 3dfx cards from the same manufacturer to run SLI mode. Here is a quote from the Voodoo2 FAQ at 3Dfx. What are the requirements for Scan Line Interleaving (SLI)? Voodoo2 requires two open PCI cards; two identical Voodoo2 based cards from the same manufacturer, and Voodoo2 drivers. Scan Line Interleaving is limited to two cards. But I noticed when I swapped out one Voodoo2 8 MB board, and put in another 8 MB Voodoo2 from another manufacturer, Windows didn't prompt me for new drivers, it didn't even recognize that there was a new piece of hardware in the system, it thought it was the exact same board. This got me thinking, if Windows thinks that they are the same board, than I should be able to stick another board, and it will think it's the same too. I tried it, it works. Windows detected another "Voodoo2 3D Accelerator", and detected SLI mode in the 3dfx control panel of my Display Settings. I tested this by going into Quake II and trying to run 1024x768 on the 3dfx display driver, and it worked. Now just as another test, this morning I threw in a 12 MB Voodoo2, along with an 8 MB Voodoo2, to test this again, I thought there was no way this would work. Windows detected the board, but it seems to think that this is an 8 MB board, maybe because it's SLI'd with another 8 MB board. So, Windows seems to think that my 12/8 configuration is an 8/8 setup, but it still detects SLI mode, and let's me play at 1024x768. I have yet to try a 12/12 configuration with different manufacturers boards, which I am going to try later today, but I'm guessing it will work if an 8/8 configuration works. Now what did I have to do to accomplish this? Easy!, just use the 3Dfx reference Voodoo2 driver on both of the boards (which you can get here), overwrite your current drivers, and install the reference ones from 3dfx. I'm not certain if this will work on boards that aren't made on the 3dfx reference design, like the Canopus Pure 3D, and the Miro Highscore, but they *should* work. Please email me if you try this on different board configurations and have success or failure.



WOw :o :o

I`d like to buy a Voodoo2 running at 125 Mhz  :-[ but my V2 is at 90 MHz and 100 Mhz Memory, like all V2`s  :-/

As it sounds, this card must be extremely rare, if it was ever produced. What I seriously wonder is if this card was launched just before the Banshee or before V3. If it was so, then it is no wonder that this card appeared and disappeared so quickly, as even running at 125 MHz, the V2 would be seriously limited by a number of factors (mainly geom), compared to a V3, and in some cases, even by a Banshee.

wrote on 19.03.04 at 09:33:26:

What I seriously wonder is if this card was launched just before the Banshee or before V3. If it was so, then it is no wonder that this card appeared and disappeared so quickly, as even running at 125 MHz, the V2 would be seriously limited by a number of factors (mainly geom) compared to a V3, and in some cases, even by a Banshee.

This Voodoo2, if SLI'ed, would have beaten the sh!t out of the V3. D@mn, I can't remember what CPUs were available by the time V3 arrived. I doubt there was any CPU able to max a V3 out back in those days. But then again, I might be wrong (memory leaks  ;D).

Hm, I remember quite well those days :

The intel Pentium 350 & 400MHz and the mighty BX-Chipset were just released by intel, hard to get though, and at insane prices.

Pentium 333MHz was considered High End on an LX Chipset in those days.

My first Voodoo2 SLI was initially running on an AMD K6-233, and days later on an AMD K6-266 (before trying the intel P2-333, since the K6 didn't really push the Voodoo2 SLI :-/ )

I'm sure that the 125MHz boards weren't ever released in any form.  Shame really, but then the Voodoo2 was CPU limited at the time anyway!  As FalconFly says, you were lucky if you could afford a 300MHz+ system, and Voodoo2s weren't maxed out until systems well over twice that fast were released.

As for the mismatching cards in SLI.  Many combinations do just work.  And for those that don't there are special drivers to try for Windows 95/98 and ME users.

I think we could even go beyond 125mhz!!
because the limit is the memory.
The 100mhz limit is at 118/117 , based on that the limit of this memory would lay around 17%.
That would make a whopping 146MHZ!!
(people just speculating)
And a 146mhz V2 SLI would have a filrate (theoritically)of: 292mpixels/s Single
and: 584 mpixels/s Multi

That  584 Would KIL a VOODOO4 (maybe)

When you are putting the problem of using higher GPU clock, you must remember the old problem:
1) A faster GPU offers you better performance, as much as the fill-rate is of concern, and overall giving a better performance at high resolutions. However, in some games you could not run the Voodoo2 at 1024x768 resolution because you have only 12 MB of textures. Actually, the faster GPU would bring not too much of a performance increase.

2) The actual speed would not be greately improved because the GPU has to wait the same amount o time when he needs to get textures, or be issued commands. Thus, most of the time, a faster video memory, is a greater performance boost, rather than a GPU clock speed increase.

We all know that the V2-SLI combination was almost on par with the Voodoo3 -2000. This does not mean that if the V2-SLI's GPU's were clocked faster the performace increase would be so big. And also don't forget that the V2-SLI can have around 12 Mb for textures, while the Voodoo3 (not the 3500) can have upto 14 Mb for textures(video mode is using around 1Mbx2 double buffering reasons) on the video memory.

The last point is that we all should know by now that the V5-6000 performance was not fully unleashed, because there were faster memory chips required to use at maximum the 4 powerfull VSA-100 chips.

To figure this results for yourself, you have to overclock the card by changing only the GPU clock speed (some can change it at the same time with the memory clock and this is not relevant for the memory limiting), and see the speed increase in games. You will soon find out that it is not such a big increase.

Ehh.... correct me if I'm wrong but I thought the V2 SLI made of 2 12mb voodoo2's  had 16mb texture memory (8mb on each card). And The memory bandwith of a Voodoo2 is everything far from little 192bit wide with a bandwith of 2,1gb/s in SLI that is 4,3gb/s.

Well we have a problem with the Voodoo2 aint it.
We can only overclock GPU AND Memory. Unfortunately there is til now no way to overclock only the memory.
But Can someone tell me if that is even possible????

Reffering to bandwidth: the value you mentioned is the theoretic maximum, or the peak value if ever achieved in reality. Don't forget that PCI is still PCI. You can't have too much bandwidth even if you use 2. Or at least you can't have twice the speed, but somewhere around 1.5, 1.8 the speed. 2 PCI card use 2 slots, 2 IRQ's, 2 ... At a moment, the PCI latency's becames a major setback in case of multiple PCI cards.

Overclocking solely depends on the videocard design. Most of the newer cards have independent clocks for GPU and memory (asyncronous), but older ones have a single clock that controls both the GPU and the memory, even if they run at different speeds.

The V2 is not having async clocks management, so overclocking only the  memory is almost impossible (unless you modify some circuitry on the board itself).

Well I would modify the board but nowone that I know, knows how to do that.  >:(

He the problem with the Voodoo2 is the PCI bus I know, so to solve this we should implement HSR into Glide.
HSR Hidden Surface Removel. But I don't know if  Dborca allready implemented it into Glide, because there was some release with HSR but I thought it was some other technique or or ment 'H' Version Second Release.
But they are Using HSR removal in Nvidia drivers etc.
And I could notice it with 3Dmark99 with game1 race the tunnel part, with my gf4 I get a performance increase at that part, with my voodoo2 I get a drop in performance. Conclusion HSR would be very useful!!!

Hmm...true but it wouldn't be easy to implement it in GLiDE! ::)

And if you read the threads 'carefully' you'll see that Daniel is the genius who is working on the Voodoo Mesa drivers - not GLiDE! ::)

Still, a nice idea! ;)

Yes but Opengl converts it to GLIDE3x.
And yes I am thinking about it, but directx is also one of my interests to wrap it to glide.

I took the example of the Voodoo5, where you don't have 64 Mb of Video Ram, but somwhere between 32 and 64 actual memory (depending on situation). If Voodoo2 is different, very well then...

Regarding the resolution limiting, I knew that under normal conditions you couldn't reach 1024x768, we don't even talk for a Voodoo2 about using Tripple buffering...

Anyway, having a Voodoo2 at 125Mhz GPU (even if we increase the speed of the memory) we don't get really too much of an improvement. Always, unless the workload is too big, a Banshee outperforms clearly in terms of quality the Voodoo2. And I presume most people are choosing 3dfx for quality, right?

Well, with such old Card running slightly more recent Games, the 125MHz sure would yield almost a linear acceleration.

(remember, most people run well above 1000Mhz systems, that put a normal 90MHz Voodoo2 absolutely on its limits)

The Voodoo2 IMHO will always have one big advantage over the Banshee : Single pass Multitexturing
...which yields a massive performance boost, as well as a great image quality compared to single-textured scenes ;)

I know about that multi-texturing advantage. In fact 2 textures per pass, since this multitexturing word is a bit too general. The massive performance boost: probably in SLI mode. Otherwise the boost is not so big.

I remember seeing a Savage 4 (110 GPU/ 125 Memory) running with and without Multitex, with and without texture compression (Quake3 in 16 bit color max details). The card was having an advantage on 1024x768 for texture compression (3-4fps more) otherwise the advantage is lower at 800 or 640 resolutions.

Regarding multi-texturing, the gain was almost similar in 1024 resolution as when the texture compression being used (3-4 fps more). In 640 and 800 the difference was less visible. Even without those two options, the card was still able to do more than 20 fps in 1024x768x16.

And the Banshee (110 GPU/ 125 Memory) was able to do a bit less than Savage 4 with multitexturing (no texture compression), but still above 20 fps. Multitexturing benefits varies upon situation, but it gives a certain 2-4 fps increase in most situations.

Quality is more of a concern for non-single pass multitexturing cards mostly when you need bump-mapping (emboss most of the time). Then, a non-Mtex card is showing less quality, even when it is forced to do the effect as precise as possible. But other than Slave Zero and few other games, bump-mapping is kind of rare for 3D games. This is why non-Mtex is actually not such a big loss in quality most of the time, but a certain loss in speed (gets bigger as more texture are required for an object).

*ugh*, the boost is nothing less than upto the theoretical ~100% for scene-wide multitexturing, and was observed upto ~80% in real games (?)

Even Quake2, which only used a fraction of Multitexturing effects more modern Games use, yielded a significant performance boost, allowing it to outpace the higher clocked Banshee.

You shouldn't compare Oranges and Apples (Banshee vs. Savage4, which are 2 entirely different architectures), when the item of interest is about Banshee vs. Voodoo2 ;)

PS.
There is no such thing as "Multitexturing gives x-y fps more". It can be close to +1 fps on a completely bandwidth starved Card, while it could be +50fps on a modern Architecture forced into several rendering passes (and we're not even talking different CPU's involved yet).

I admit that the comparison might not be the best with Savage 4 and Banshee. But Savage with and without multi-texture is at least a fair proof of what you are most likely to get, and it is not so much. A possible response to this question is that you can never forget that the CPU and PCI bus are limiting your performance at first, and the actual software code is the biggest difference sometimes. And finally there are games and games. Quake is not a good example of multi-tex. Apart from some shadowmaps and some lightmaps, the whole game is kind of single textured objects. Sometimes they never use Mtex, other times they use it heavily (Unreal Tournament).

Quake 2 is a different problem. First of all Quake 2 normally used a different path for the OpenGL wrapper (had a special one for a Quake 1 and a special one for Quake 2 as opposed to the only full OpenGL ICD available for Banshee). Remember the fxmemmap.vxd file in Quake2 directory? Not to be forgotten is that on a Banshee the GPU must also refresh the 2D display frames at the same time, while the Voodoo2 lets this job on the 2D card. The speed hit can be calculated on paper, and it is not indiferent for the performance.

Also, a test on Quake 3 with an OpenGL for a Voodoo2 (and Quake3 only) proved that the Banshee could get a few more fps. So again, comparing Banshee and Voodoo2 and solely saying that a big performance is gained by the usage of Mtex is not founded.

Oh well...

Although I don't like using Benchmarks other people did, here's what I deem reliable enough to be representative:

http://www.falconfly.de/temp/banshee-q2.gif http://www.falconfly.de/temp/voodoo2-q2.gif
( Source : THG )

As you can see, the Multitexturing advantage is alot bigger than you think Boui.
The Voodoo2 is a clear winner against a considerably higher clocked Voodoo Banshee.

Yes and that is only with a Voodoo2 consider what to voodoo2's can do!

PS. if the voodoo2 is used in SLI. Does it than use 2 PCI busses ( I thought so). Or does all the DATA goed trough the SLI cable. Probably not, but maybe I am wrong. But also a banshee only uses AGP1 64bit 66mhz.
If you have 2 PCI slots with 64bit and 66mhz you have the same bandwith as a banshee!!!!!!
The Banshee can't compete with that.
And yes most V2 SLI users use them allongside fast Geforces/Radeons/V5's, and those PC's have (like I do) 2ghz+.

As far as I know, the Voodoo2 assigned Slave in an SLI setup only draws power from the PCI Bus, with all its rendering tasks taken from the Master via the SLI Cable only.

I think, the only 3D AddOn-Card existing to really rely solely on PCI BusMastering for all operations, was the infamous VideoLogic PowerVR Card (e.g. Matrox M3D).

Since the Voodoo2 is only a normal 32bit PCI Card, the 1x AGP Banshee (rather PCI66, or AGP 0x as ATI would name it for example) has an advantage loading up new Textures.

That's mostly a theoretical advantage really, but at least the Voodoo3/4/5 AGP vs. PCI often scored a small lead using the faster Interface.

If the cards were to do nothing but shoveling excessive Geometry Data or Textures over their respective Bus all the time, the advantage could be significant, close to 100% ;)

Falcongly (why you spell my name wrong?) I have to add the following: I am comparing a Voodoo2 (90/90)with a GA-630 Voodoo Banshee (110 Mhz GPU, 125 Mhz Memory).

I haven't figured out what test results you showed in here. It was a good thing to add the link to the test. The first Banshees are normally running at 100Mhz and have 8Mb video memory. Voodoo2 runs at 90Mhz and has 12Mb video memory. The fastest Banshees ran at 125Mhz and have 16Mb video memory. These improvements make a big difference, and the whole test with Quake2 is almost sensless. Quake 2 was playable (above 25fps on a Voodoo Banshee in 1024x768 or around 30 fps in 800x600 with no 3DNow! Quake). Voodoo2 normal was somewhere around 40-50fps. Actually not such a big difference.

I would refrain on posting this kind of test results like you did, without saying what were the full hardware specifications of the Video card at least. And, also, a crappy motherboard can totally screw up a test. The Intel BX400 (I guess) was a very bad P2 motherboard, at least from a certain manufacturer: DTK.

And when you think that you play a game usually at 1024x768 (when it performs acceptable), the Banshee becomes more than attractive compared to a normal Voodoo2. And the 16Mb texture memory start to make a difference in games better than Quake2, even if the card gets a slight performance hit when using multi-pass single-texturing. Most of the time is pretty useless what performance you get in 640x480 resolution, since most of the time you wouldn't use the card in that resolution.

Always the Banshee was put at tests against Voodoo2 in SLI and in not so great games such as Quake2. Quake3 really pushed the limits of the Voodoo2, even in SLI. If the game is not stressing too much the cards, you don't see some of the benefits of the Banshee. Also, when Banshee can usually outrun a Savage4 with single-pass multitexturing running at the same frequency, you draw the conclusion that Voodoo Banshee is a very good card, and in some respects clearly better than the Voodoo2.

Sry for the misspelling.

Anyway, the framerates you are talking about were already exceeded by the Voodoo1, and even at that time, people liked to stay well above 20fps (?)

So you'll hardly find people who were used to run a Game at 1024x768 on a Banshee (since this puts the Card already on its fillrate limit).

I agree, on more modern Games than Quake2, the 16MB do have an advantage, but then, more Multitexturing automatically comes to play.

If you read the numbers correctly, you will see that even a 125MHz Diamond Monster Fusion would have problems reaching the Single 90MHz Voodoo2.
Even if the Board used wasn't the biggest and greatest intel BX Platform to ever exist : since both Cards ran on the same, it makes no difference at all.

And I totally disagree in one point :
Quake 2 was and is an excellent comparison for old Cards. There are not many Games that natively supported Software, Glide, Direct3D and OpenGL Rendering.

And this one is beyond me :

Quote:

Quake 2 was playable (above 25fps on a Voodoo Banshee in 1024x768 or around 30 fps in 800x600 with no 3DNow! Quake). Voodoo2 normal was somewhere around 40-50fps. Actually not such a big difference.

So the jump from 30fps to 40-50fps is "not such a big difference" ?
You're contradicting yourself completely, calling a ~25-33% Framerate advantage for the lower clocked Voodoo2 "not such a big difference", especially in that context  ::)

If you "benchmark" Cards on "playable", you certainly can't call your work "benchmarking", whatever you were doing there.


Quote:

Also, when Banshee can usually outrun a Savage4 with single-pass multitexturing running at the same frequency, you draw the conclusion that Voodoo Banshee is a very good card, and in some respects clearly better than the Voodoo2.

Nono, you draw the wrong conclusion there (remember that "Apples vs. Oranges" thingy). I'd rather say the Savage4 really never was 'king of the hill', and offers inferior performance when put to comparison with the 3dfx products of its time.
Sure, the Banshee is a very good Card, but (if you remember) we were talking about performance.
There, the Banshee can truly excel only in Single-Textured Games... (that was my whole point)

PS.
I find it somewhat funny trying to explain to you things, that an entire 3dfx Community was cristal clear about already 7 years ago ::)

Quake2. The first thing that comes to my memory is the special Openg32.dll released by ID (with 3dfx support). This is not a full OpenGL ICD. Quake 1 had this kind of OpenGL to Glide wrapper. These so-called miniGL had been long time optimised for Voodoo1 and Voodoo2. For Voodoo Banshee, 3dfx was working on a full ICD at that time. I remember the poor fix for NFS3. It looked horrible in Glide when this patch was applied. What I want to point is that when Banshee appeared, a long time it didn't had the best support as oposed to Voodoo2. Especially games as Quake2 that were designed from the start for Voodoo2.

Full OpenGL ICD. I remember seeing a Banshee running with the latest Quake 3 OpenGL for Voodoo2. Apart from the problem of multi-texturing, there was a certain improvement, sometimes of more than 5 fps in 800x600 resolution, as oposed to Banshee full OpenGL ICD. And don't forget miniGL is based on glide2x (smaller and faster for simple jobs), the full ICD is based on glide3x. This can make for some subtile differences. Testing both cards in Glide2x on a game would be the best compromise in seeing how they handle the load.

Feel free to use the most current and optimized Drivers for a comparison, and post the Results.

But for now, all you've been giving in Results have been the opposite of what you were claiming (?)

You could use Quake 1 (Banshee friendly) vs. Quake 3, Unreal 1 and Unreal Tournament (Multitexturing and more Texture heavy).

I have pointed in one direction. In certain aspects Banshee is better than Voodoo2. Even if it lack the 2 textures per pass, the real situations does not depend so much on this. Ok, you get 10, 15 fps. But again, using a low resolution. Also, to be noted is the fact that quality is lower on a Voodoo2 as oppsed to a Banshee.

The whole point about multitexturing and it's impact on performance was this: you certainly get a fairly good improvement, depending on the situation, but it is not solely the reason for which the Voodoo2 could outperform the Banshee. Until Voodoo3 arived there was no card to beat up the performance of a Voodoo2 in SLI. And this clearly proves some excelent performance capabilities from the Voodoo2.

And getting a whole discussion back, a Voodoo2 would not offer too much of an improvement if the GPU would be overclocked, unless the memory response would be also fasten up.

Quote:

And getting a whole discussion back, a Voodoo2 would not offer too much of an improvement if the GPU would be overclocked, unless the memory response would be also fasten up.

In case you haven't noticed :

On the Voodoo2, there is no independent Memory/GPU timing...
Overclocking the Voodoo2, you automatically overclock the Video RAM as well ::)

I know there is one clock for GPU and Memory on Voodoo2. I said this earlier. The whole problem was what would happen if the GPU would be clocked faster as opposed to what would happen if the Memory would be faster. Pure theory...

You're the first one to ever claim that (?)

So what are the SSTV2_ Commands for separate GPU/MEM Clock ?

Can you do that clock change for any layer? D3d/OpenGL/Glide. I might be wrong, but I remember that only in Glide/OpenGL you can do that. Sorry if I am wrong. I know little about this settings effect.

wrote on 14.05.04 at 13:27:34:

You're the first one to ever claim that (?) So what are the SSTV2_ Commands for separate GPU/MEM Clock ?

Tom's Hardware mentioned it in an old '98 vintage article but I think we can discount that as being incorrect.