WiiGeePeeYou (Hollywood) what IS it ?

Status
Not open for further replies.
pc999 said:
Well if wee are going to suposse 9800 we already have a idea from what is possible to do it that +P4 2,5Ghz (?) (in the same league as a AXP2400-3000), anyone remember the real time first Heavenly Sword video back in 2004.



That s hard to say specially because each one count polys in their own way, but search for ERP post on the subejct in other wii/rev threads that can give you more info.

No i have an actual benchmark showing that to be the case, 7.6 million polys with 8 lights, and 18million with 1 light, actual rendered polys for the GF4 Ti4200, which had more vertex optimizations than xbox1 had....
gf4ti4200bench.JPG



tell me that isnt as damn close to the xbox as youve ever seen in a benchmark? (btw the compared system is an athlon64 3000+ with a 6800GT @ 432//1140 =)
 
Last edited by a moderator:
Yes but Factor5 may count it in other way, after all GC only do 12M, BTW thanks for the graph although remember that XB is a closed platform.

BTW anyone knoes here I find the first trailer for HS?
 
pc999 said:
Yes but Factor5 may count it in other way, after all GC only do 12M, BTW thanks for the graph although remember that XB is a closed platform.

BTW anyone knoes here I find the first trailer for HS?

no, nintendo claims 6 - 12 million real world, factor 5 surpassed that at launch (15 - 18m for rogue leader). xbox is a closed platform true, but the Ti4200 had dedicated memory which was faster than the memory in the xbox1, it had vertex optimizations th xbox1 didnt. and it didnt texture in such an ass backwards way as to render any advantages of a closed box useless (apparently it grabs portions of a texture on a per poly basis? doing hundreds of thousands of fetches on high poly games?)

And regardless of how they count the polys, they are running 8 hardware lights, and it easily is more than 7.6 million polys, even at the worst case.....
 
Last edited by a moderator:
Wasn't flipper more or less an 8 pipeline chip, so it's poly count isn't halved when doing 8 lights? (isn't Xbox's lighting higher quality anyway though?)

Plus, flipper was fixed function, I believe NV2A had a fixed function mode as well, not to mention it may have been possible to have the cpu aid in the calculations.
 
chosen_colette said:
This is true, now heres a question for everyone here, is it at all possible that the xbox1 cannot keep up with the gamecube when 8 hardware lights are being used? For example Rebel Strike (21million polys/sec with 8 lights, 8 effects).
Did Rebel Strike use 8 lights? I know it used 8 textures, but I didn't think it used 8 lights.
 
mattcoz said:
Did Rebel Strike use 8 lights? I know it used 8 textures, but I didn't think it used 8 lights.

you had the light from the environment which i might add also used light scattering, you had dozens of ties all shooting at you at any given time, each laser beam was a light source, so yes i believe it was pushing 8 lights...
 
Fox5 said:
Wasn't flipper more or less an 8 pipeline chip, so it's poly count isn't halved when doing 8 lights? (isn't Xbox's lighting higher quality anyway though?)

Plus, flipper was fixed function, I believe NV2A had a fixed function mode as well, not to mention it may have been possible to have the cpu aid in the calculations.

Flipper was a 4 pipeline chip actually, and im not exactly sure how polycount is affected on cube, it doesnt seem to be at all given rebel strike.... as for the xbox, the self shadowing in rebel strike wouldve had be have used the vertex shaders on xbox, taking even more poly pushing power away from the machine....

Not to mention that any xbox game with per pixel lighting also had an abysmal polygon count and framerate (halo2, thief 3, splinter cell chaos theory, riddick anyone?) with per pixel lighting xbox cant even push 7 million polys/sec, that figure in that benchmark was lit per vertex, just like cube does its lighting in the majority of cases...
 
Last edited by a moderator:
chosen_colette said:
Flipper was a 4 pipeline chip actually, and im not exactly sure how polycount is affected on cube, it doesnt seem to be at all given rebel strike.... as for the xbox, the self shadowing in rebel strike wouldve had be have used the vertex shaders on xbox, taking even more poly pushing power away from the machine....

Not to mention that any xbox game with per pixel lighting also had an abysmal polygon count and framerate (halo2, thief 3, splinter cell chaos theory, riddick anyone?) with per pixel lighting xbox cant even push 7 million polys/sec, that figure in that benchmark was lit per vertex, just like cube does its lighting in the majority of cases...

Hmm, I thought that Flipper may have been like the X1x00 series, where the elements of the 'traditional' pipeline seem largely decoupled from each other.
Only 4 pixel pipelines, but 8....I dunno what you would call them, ROPs?
 
Fox5 said:
Wasn't flipper more or less an 8 pipeline chip, so it's poly count isn't halved when doing 8 lights? (isn't Xbox's lighting higher quality anyway though?)

Plus, flipper was fixed function, I believe NV2A had a fixed function mode as well, not to mention it may have been possible to have the cpu aid in the calculations.

Arent you talking about the TEV that read 8 textures?
 
Fox5 said:
Hmm, I thought that Flipper may have been like the X1x00 series, where the elements of the 'traditional' pipeline seem largely decoupled from each other.
Only 4 pixel pipelines, but 8....I dunno what you would call them, ROPs?

from what i understand. Flipper is more like a pre-shader-era GPU. differences are,
embedded memory, and the TEV decoupled from the pixel pipelines, instead of TMUs in each pixel pipeline. but Flipper has only 4 ROPs, AFAIK.
 
Fox5 said:
Hmm, I thought that Flipper may have been like the X1x00 series, where the elements of the 'traditional' pipeline seem largely decoupled from each other.
Only 4 pixel pipelines, but 8....I dunno what you would call them, ROPs?


Fox, it does not make sense to have more ROPs than fragments you can produce per clock.

btw, the classic rasterization pipeline has samplers, combiners and ROPs. but none of those has anything to do with vertex lighting - that's not in the rasterizer but in the TnL part of the rendition pipeline.
 
chosen_colette said:
This is true, now heres a question for everyone here, is it at all possible that the xbox1 cannot keep up with the gamecube when 8 hardware lights are being used? For example Rebel Strike (21million polys/sec with 8 lights, 8 effects). My evidence shows that a GF4 Ti4200 (similar to the xbox but even more powerful) can only push 7.6million polys with 8 lights, and thats not taking into account the vertex shaders in xbox wouldve had to been doing the self shadowing rebel strike used also..... Whats your guys take on that?

That's utter nonsense. Rebel Strike didn't push 21mill polygons/sec + 8 HW lights + 8 textures.
We've had this discussion before here and in a million other threads I am too lazy to dig up. Here's what the developers says
ERP said:
Short of a hardware analyser I would be extremly suspicious of any polygons/second numbers. I couldn't guess accurately and I doubt anyone elses ability to do so.

I would bet that neither RE4 or Starfox are anywhere near 12 million polygons/second.

Most in game polygon counts are a LOT lower than most of the guesses I see around and you'd be surprised just how much you have to increase polygon counts to get a visible difference.

Factor5 is counting culled polygons, etc. in that number. 21mill p/s, 8 HW lights, 8 textures would even be beyond the theoretical limits of Flipper, go figure!

http://www.segatech.com/gamecube/overview/index.html
1 vertex color + 1 light + 1 texture 20M polygons/sec
no vertex color + 1 texture 26.4M polygons/sec
1 vertex color + no texture (gouraud shading) 32M polygons/sec
 
pc999 said:
Originally Posted by hupfinsgack
I'll consider it to be fake until there's confirmation or visual evidence, i.e. until we see some games showing improved graphics.

BTW, wouldn't a GPU with 8 pipelines be a hindrance for normal mapping intensive engines (UE3)?
Number of pipelines without the speed is useless.

Well, how about 8 pipelines @ 243 Mhz (the rumoured clockspeed)?
 
chosen_colette said:
you had the light from the environment which i might add also used light scattering
The fabled light scattering is some minor TEV texture faking/trickery (at best; could also be totally pre-baked textures I suppose), completely separate from the hardware lights supported by the T&L unit.

you had dozens of ties all shooting at you at any given time, each laser beam was a light source, so yes i believe it was pushing 8 lights...
I'm doubtful any polygon actually had 8 lights acting on it ALL AT ONCE. If even a majority of polys on-screen had, the scene would be bright as daylight. Except in tie fighter laser-green.

I'd say 1-4 lights on any one polygon at most. IE: directional light (the sun), possibly an ambient light, and 1-2 laser blasts. Most polys would only have 1-2 sources on it - the directional/ambient light(s), as the range of the laser blasts isn't super huge.
 
Guden Oden said:
The fabled light scattering is some minor TEV texture faking/trickery (at best; could also be totally pre-baked textures I suppose), completely separate from the hardware lights supported by the T&L unit.


I'm doubtful any polygon actually had 8 lights acting on it ALL AT ONCE. If even a majority of polys on-screen had, the scene would be bright as daylight. Except in tie fighter laser-green.

I'd say 1-4 lights on any one polygon at most. IE: directional light (the sun), possibly an ambient light, and 1-2 laser blasts. Most polys would only have 1-2 sources on it - the directional/ambient light(s), as the range of the laser blasts isn't super huge.


Well based on the benchmark i provided of a similar GPU, witth 5 - 8 effects and even 2 - 4 lights, Xbox still couldnt push enough polys in that situation to do rebel strike, as its throughput would be halved with the second pass required for more than 4 effects...
 
http://biz.gamedaily.com/industry/feature/?id=12859

John Swinimer, Senior Public Relations Manager of Consumer Products, emphasized that the Wii architecture is capable of producing far better results than what we've witnessed thus far. "I think what you saw [on Wii] was just the tip of the iceberg of what the Hollywood chip can bring to the Nintendo Wii," he said.

Dunno if this is new to anyone. But perhaps the GPU is indeed more powerfull than expected, only devs got pre-alpha kits and thus pre-alpha kit graphics in time for E3?
 
The shader just has title now,

Planet GameCube: In a recent IGNinsider article, Greg Buchner revealed that Flipper can do some unique things because of the ways that the different texture layers can interact. Can you elaborate on this feature? Have you used it? Do you know if the effects it allows are reproducible on other architectures (at decent framerates)?

Julian Eggebrecht: He was probably referring to the TEV pipeline. Imagine it like an elaborate switchboard that makes the wildest combinations of textures and materials possible. The TEV pipeline combines up to 8 textures in up to 16 stages in one go. Each stage can apply a multitude of functions to the texture - obvious examples of what you do with the TEV stages would be bump-mapping or cel-shading. The TEV pipeline is completely under programmer control, so the more time you spend on writing elaborate shaders for it, the more effects you can achieve. We just used the obvious effects in Rogue Leader with the targeting computer and the volumetric fog variations being the most unusual usage of TEV. In a second generation game we’ll obviously focus on more complicated applications.

A hardware-accelerated recirculating programmable texture blender/shader arrangement circulates computed color and alpha data over multiple texture blending/shading cycles (stages) to provide multi-texturing and other effects. Up to sixteen independently programmable consecutive stages, forming a chain of blending operations, are supported for applying multiple textures to a single object in a single rendering pass.

For example the R420 architecture uses quad 64-bit memory controllers and a 16-stage pixel pipeline which can be scaled from 4 to 16 stages in steps of 4.

Could someone explain the differences in R420 16 stage pixel pipeline and Wii(GC) 16 stage Recircular Shader, or are they the same?
 
Ooh-videogames said:
Could someone explain the differences in R420 16 stage pixel pipeline and Wii(GC) 16 stage Recircular Shader, or are they the same?

there's some misunderstanding here - a sm2 gpu does not have stages - that's a characteristic of the hardwired rasterizer, and stands for cascaded combiners.

r420 can do 16 samples per pass (i.e. per fragment shader routine). flipper can do 8 of the same in a pass. that is as direct parallel as you can do between those two.
 
I don't know how reliable it is but this supposed article from the new issue of Gameinformer seems to coincide with recent "developments".

http://forums.nintendo.com/nintendo...sage.id=1260806&view=by_date_ascending&page=1

It confirms support for normal mapping, bump mapping, HDR lighting among other things. It also talks about the dedicated physics hardware as well. On their own none of these articles seem to be much evidence of much but together they all seem to be painting a very similar picture.
 
Status
Not open for further replies.
Back
Top