PS 2.0 review I would like to see

[nRollo]

So far it looks bad for the nV3X line on the DX9/PS 2.0 shader front.

What would be interesting to me is to see the difference between the PS 1.X and 2.X screenshots showing the 2.0 effects and the benefit of them.

It's pretty hard to know if an effect is worth having when you don't know what it would do for the gameplay.
(I'm thinking back to the beginnings of EBM, when I used to argue that the V5 didn't have EBM, but the only game that really did was Giants, so whether you had it or not really didn't matter too much.}

Anyway, I'm wondering if this is more of that sort of thing: companies like to point out their competitive advantages (e.g. ATI pointing out the nV35 isn't an 8 pipeline card, when for the most part 8X1, or 4X2, equals pretty much the same thing in games) when it often doesn't matter much at all.
So, it would be good to see if the apparently huge difference in ATI PS2.0 performance means anything in practice, or is just a selling point.

 

[McArra]

yep

 

[BFG10K]

(e.g. ATI pointing out the nV35 isn't an 8 pipeline card, when for the most part 8X1, or 4X2, equals pretty much the same thing in games) when it often doesn't matter much at all.

8x1 is not the same as 4x2. It can be but in most cases 8x1 is far superior.

 

[nRollo]

It can be but in most cases 8x1 is far superior

Sure thing BFG. I've never seen a benchmark where the 5900 Ultra is even remotely close to a R9700.

 

The Radeon runs single textures faster than the Nvidia. And Nvidia runs multitextures faster than ATI.
Now as for the pixel shaders. PS 1.4 and PS 2.0 are mathematically identical. Its just that 2.0 shades in a single pass as opposed to "i would gues" two passes? Dunno for sure. But it does not sound to me like they are so different other than the amount of passes it takes to complete a shade.
GM

 

[Genx87]

No idea bout the differences between PS 2.0 and 1.4. I believe 1.4 was also could do shader operations in single pass which was one of Nvidias main arguments about 3dmark03 as the GF4 only supported upto PS 1.3 and not "that" many games really used PS 1.4.

Anywho I was under the impression Multi-texturing is going to be used more and more in favor of single texturing. I could of course be wrong on that one.

All I can say is in SWG I am forcing an override to use PS 2.0 instead of 1.4\1.1 and there is indeed a difference in performance. ~5-10% slower for PS2.0. Now I am not sure what exactly what or how the game is using PS2.0. Because by all indications the game looks exactly the same using any of the PS versions. 1.4\1.1 yield the exact same performance but PS2.0 is a bit slower. But the picture is no different between the two.

So there may be something to this but the question now that needs to be asked. What does it matter? I didnt notice a single difference in the visuals.

 

[Corporate Thug]

i gotta question: will DX 9 games only support one type of PS (ie ps2.0) or would you have the option of what version ps you want to run?

 

[Rogodin2]

"Sure thing BFG. I've never seen a benchmark where the 5900 Ultra is even remotely close to a 9700."

It isn't when full DX 9 shaders are implemented. PS2.0 and opengl 58.9e

rogo

 

[nRollo]

It isn't when full DX 9 shaders are implemented. PS2.0 and opengl 58.9e

Rogo, see above responses that say the picture looks the same between PS2 and PS1.4. If you can run the games in either, and the picture looks the same, this would be a pretty meaningless victory for ATI as the 5900s are as fast at 1.4.
That's what I was trying to get at with this thread.

 

[Rogodin2]

Got it.

I believe you will have to see it ingame for a decent comparison though 😉

rogo