View Full Version : The TEV on the Gamecube
I've been hearing a lot about this, but no one has any specifics.
From what I've heard, it sounds like a slightly more flexible version of the NSR on the GeForce 2 chips. But, at the same time, I was just told by someone who seems to know what he's talking about (but could be a kid for all I know), say that the TEV is much more flexible than the pixelshaders on the NV2A. Instinct tells me that's incorrect, simply due to the transistor count. ;)
Does anyone know? Can anyone tell? :)
Ozymandis
15-Jul-2002, 19:35
I'm curious too. How does the capabilities of the Cube's TEV compare with the nV2A's pixel shader?
From what I understand, there are a few things that can be done with the TEV that can't be done on the NV2A, however, overall the NV2A is more feature rich.
The major difference is that NV2A does all of it's texture reads first, so you can't do a deffered read based on a calculated result (unless it's one of the calculations supported by the texture stage) without resorting to multipass.
In the TEV texture fetches are interleaved with the combiner operations(though not totally freely), so you can theoretically do arbitrary deffered lookups.
However, in most ways flippers texture lookups are more limited as are it's combiners flexibility.
IMO Flipper probably has the better theoretic approach, with the downside of some potential texture cache performance issues.
Realistically there are operations you can do on NV2A that youy can't on Flipper and viceversa.
Could you give some examples of effects that are easily done on the TEV and not on the NV2A, and vice versa?
Anything that requires you to compute f(g(x)) where g(x) requires combiner stages to compute. and f(x) is computed as a texture lookup.
It quite honestly doesn't come up very often, but it's nice to have the flexibility.
Hey Erp, Started job huntinng yet? I have a question about Starfox. I remember the 'fur shading' subject coming up before Rare had even showed the pics of Fox' fur. And everyone said that the Flipper just wouldnt be able to do it in realtime (only in cutscenes) and only the Xbox Pixel Shaders might be able to do it in game. And now we've learned that Starfox' fur is in realtime AND around 60 fps with some REALLY big outdoor areas (and grass blades too). What routine do you think Rare are implementing? And what else do you think we'll see from the XB and GCN as far as grass and fur?
Goldni: I'm wondering about that too.
People compare the Flipper to GF2 and other DX7-level graphics renderers.. but it seems to be doing a lot more than those boards could do. Personally, I don't see how th' TEV could be more flexible than the NV2A's pixel-shading.. but hey, I'm not a developer. I know that Factor 5 said that the GCN and Xbox are equal in power overall, but they're a tad bit biased..
DeathKnight
17-Jul-2002, 03:02
What routine do you think Rare are implementing?
Probably just a bunch of layered textures. If you get up close to the fur on Starfox it actually looks pretty poor in quality.
I'm certain the Fur rendering technique is the same as the one used by everyone else. They're all based on the paper by Hughes Hoppes (MS Research), the papers on his homepage.
Basically you draw a number of transparent concentric shells of the model approximating a volumetric render. You can also draw fins on the model to show the fur cross-section. It doesn't require any clever shading, just fillrate.
Given how far away from Fox you are most of the time they probably need to render only one additional shell, possibly two, but I'm speculating here.
MS has demos of other variations on this technique including Grass, and volumetric trees. The trees in particular look great, but the dataset required ends up being IMO impractically large.
Steve Dave Part Deux
17-Jul-2002, 06:20
Actually, it looks best up close. In cut scenes, it's particularly sharp(duh).
DeathKnight
17-Jul-2002, 17:49
http://image.com.com/gamespot/images/2002/e32002/gamecube/starfox/starfox_screen001.jpg
http://image.com.com/gamespot/images/2001/gamecube/starfox/starfox_0102_screen017.jpg
Not too hot looking :wink:
That pic looks terrible.. I mean it doesnt even look lke a real fox.. look how big his eyes are!
And a Fox wouldnt be seen dead in that kinda gear - so outdated.
:P
Tagrineth
17-Jul-2002, 20:57
Only thing I don't like in those pics is the tail. Icky.
But that's apparently been fixed; latest screenshots in NP look *MUCH* better.
StefanS
26-Jul-2002, 08:13
Hey ERP, could please explain a little bit more about the TEVs? It is pretty easy to find some stuff about Pixel Shaders, but I don't find anything on how the TEVs exactly work and what principles they're based on. Thanks.
aaaaa00
26-Jul-2002, 10:53
Not too hot looking :wink:
He's right. Up close the illusion breaks down, and you can see the layers of transparent textures used to create the effect.
Look at left side of the picture.
Not to mention it's really easy to LOD this effect -- it's simple to avoid rendering this many fur layers when you're far away from the character.
Hey ERP, could please explain a little bit more about the TEVs? It is pretty easy to find some stuff about Pixel Shaders, but I don't find anything on how the TEVs exactly work and what principles they're based on. Thanks.
The TEV and "pixel shaders" are basically cute acronyms for what used to be called color combiners. The TEV also incorporates the Texture reading part of the pipeline.
A color combiner is in general implemented as a single logic op, in NVidia's case thats public (register combiner docs) and is of the form
A op1 B op2 C op1 D
where op1 is either Dot Product or multiply, op2 is either add or select.
As you can see by repeating this multiple times with some register manipulation between stages you can do most basic math. Pixel shaders just provide a simple consistent interface to this (and other vendors implementations).
The TEV uses a different basic combine operation which is a little more limited. However since the Texture reads can be interleaved with the combiner operations it allows you to do things that would require multipass render on NV2X.
So as an example
on NV2X I have to write
Texture Read
Texture Read
.
.
Combiner Op
Combiner Op
Combiner Op
Combiner Op
.
.
.
On Flipper I can write
Texture Read
Combiner Op
Combiner Op
Texture Read
Combiner Op
Combiner Op
Texture Read
Combiner Op
Combiner Op
.
.
.
I guess the easiest explanation is that Flipper has simpler units for combining and reading textures, but allows more complex arrangements of the units.
So if one of the texture reads is dependant on a previous combiner Op and you can't squeeze the ops into the texture addressing instructions the NV2X would require multipass to do the same thing.
Yes at least a good explanation of TEV thanks ERP,
it's really frustrating not to find any docs or in depth information relating to Flipper. TEV is a 16 stage pixel pipeline right (I think I read it a while ago in an interview of Greg Buschner..) ? basically any stage in this pipeline can be a texture read or a combiner op unlike Nvidia's Register combiners if I understand well. What kind of combiner ops the Flipper can execute ?
Luminescent
02-Dec-2002, 17:53
What kind of math combiner ops can the flipper implement on its TEV (dp3 or dp4)?
Luminescent
02-Dec-2002, 18:15
On a second note, it was previously established that the 6:6:6:6 and 8:8:8 bit color modes of the flipper are a limitation of its framebuffer. However, do its internal units (TEV, Texture coordinate generator, etc.) support a higher level of precision? Is it comparable to the NV2A (Integer 10 format)?
CaptainHowdy
02-Dec-2002, 18:27
That pic looks terrible.. I mean it doesnt even look lke a real fox.. look how big his eyes are!
And a Fox wouldnt be seen dead in that kinda gear - so outdated.
:P
only problem is, thats an old screenshot from when the fur was first shown, in the final cut, it looks flawless, the graphics in SFA are flawless, if only they had worked that hard on the gameplay.
BenSkywalker
03-Dec-2002, 01:15
only problem is, thats an old screenshot from when the fur was first shown, in the final cut, it looks flawless, the graphics in SFA are flawless, if only they had worked that hard on the gameplay.
Where did you get your copy? Mine still looks just like the early screen shots, including the very obvious and ugly alpha texture artifacts(rather serious 'clumping'/aliasing).
Hehe.. I was trying out sarcasm CaptainHowdy
Still haven't got it pinned down it seems. :wink:
EDIT: does the date of your posts show 07 Feb 2002??
Luminescent
03-Dec-2002, 17:10
How many theoritical texutre-reads/combiner ops can the flipper TEV perform per cycle? Also, for those of you who did not read my previous post:
It was previously established that the 6:6:6:6 and 8:8:8 bit color modes of the flipper are a limitation of its framebuffer. However, do its internal units (TEV, Texture coordinate generator, etc.) support a higher level of precision? Is it comparable to the NV2A (Integer 10 format)?
Legion, I did see your post, I'm not really comfortable getting this specific.
Also I'm not sure what the actual intermediate bit precision is on Flipper.
Can you be more specific about why you don't want to be more specific :) Seriously is this sort of thing supposed to be kept secret by developers or are you afraid you'll start a flame war or something?
Because I really can't see how some specifics on Flippers combiner ops can start a flame war here.
I'm not really worried about flame wars, I try and stay out of them for the most part.
Technically the NDA's cover everything about the hardware. I sometimes skirt the edges of NDA's, if the information is publically available. Conveniently NVidia describe in a lot of detail how NV20 and by inference NV2X functions. Nintendo don't do the same.
The exact function of the Flipper combiner isn't public, so I'd rather stay away from the issue. Personally I can't see any real reason the data is particularly secret, but I did sign the NDA.
When does this NDA end? When you stop deving for GameCube?
Its really annoying how secret Nintendo keeps things like this. As you say there just doesn't seem like any good reason to keep this sort of thing secret. Its almost like being secretive for the sake of it.
Heh, yeah.. there doesn't seem to be a reason for their secrecy.
Then again, outside of the game development sector.. who needs to know? I guess that's their policy..
Luminescent
03-Dec-2002, 20:45
ERP, I can understand you not being comfortable revealing specifics on the flipper. However, can you tell us if precision-wise (integer precision) it is competitive with the NV2X? Can it execute a significant amount of texture-reads/combiner ops per clock?
These are relative questions, so a yes or no would suffice.
Luminescent
05-Dec-2002, 15:16
Anyone else with interest?
Several copies of the Gamecube SDK documentation have been leaked on the internet. Hardly rampant, though.
Glonk
Any idea where I could get one?
I have no idea, I only got a look at a couple of the sections (conveniently enough mostly about the TEV) because one of my classmates printed out some of the PDF files and had them at school.
They were very light on technical details, though, I didn't learn anything I didn't already know from these forums for the most part. I was hoping they'd go into more detail on how the TEV works, rather than just how to use it.
Maybe that's in another section I didn't see...
If you could find out where your friend got it from that would be great, or even what the file was called so I could search for it myself.
megadrive0088
05-Dec-2002, 22:57
i was in a chat last year were several people who had direct knowledge of Flipper talked for well over an hour about the various blocks within Flipper including TEV. i wish i could remember the finer details, but i cant, and i'd probably get it wrong so i won't bother. basicly, the Flipper was able to do thing that even the NV2A in XBox cant do. although NV2A was known to have more features overall.
Flipper's main advantage is it's on-dia 1T-SRAM - slightly over 3MB of it.
Anyway, I am curious about what the GPU/VPU will be for Nintendo's NEXT machine. ATI wants the contract. ATI will unviel R400 in summer 2003. R500 is also being worked on, for i assume, a 2004 release. R500
will be the second major chip from the ArtX/Flipper team (R300 was 1st)
assuming R400 is being done by R200 team. I'd expect the next Nintendo to use a chip that is one or two generations beyond R500.
if R500 comes in 2004, and the next Nintendo in 2006, that's two years after R500. so... either R600 or R700 will be "Flipper2" if you can call it that. there's no telling if Nintendo's next machine will be anything like GameCube. it might be completely different. but the technology is being R&D now, or at least researched...
Who knows if Nintendo will want to foot the bill on a highend ATI part. They may prefer to go for a Radeon 9500-esque part.
PC-Engine
06-Dec-2002, 02:07
Who knows if Nintendo will want to foot the bill on a highend ATI part. They may prefer to go for a Radeon 9500-esque part.
By 2006 the 9500 core would be severely outdated and dirt cheap. I don't think Nintendo wants their next console that crippled and that cheap. :wink:
By 9500-esque I meant a cutdown/lowcost version of the current (for 2006) highend core.
megadrive0088
06-Dec-2002, 05:54
By 9500-esque I meant a cutdown/lowcost version of the current (for 2006) highend core.
possible. but i was thinking that ATI will be contracted to build a graphics part that is talored to Nintendo's desired specifications. they may start with a R700, take what's not needed, out, and put in what Nintendo wants.
Or it might be a totally new design, apart from the Radeon line.
It could be another Silicon Graphics/ArtX-esque part with some ATi influence.
PC-Engine
06-Dec-2002, 07:14
By 9500-esque I meant a cutdown/lowcost version of the current (for 2006) highend core.
possible. but i was thinking that ATI will be contracted to build a graphics part that is talored to Nintendo's desired specifications. they may start with a R700, take what's not needed, out, and put in what Nintendo wants.
Or it might be a totally new design, apart from the Radeon line.
I think the graphics chip will have embedded memory again like Flipper.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.