Game developers on Ps3.0.

[Nv40]

like if this topic was not being talked enough... :)

there is a good interview in gamersdepot where they asked the same question to a group of game Developers.. is the first part.. in a second more developers will follow soon.. :)

http://www.gamers-depot.com/interviews/dx9b/001.htm



GD: Is there ever a need to use 32-bit precision over 24-bit? I'm looking for an example where 32-bit precision shows an obvious superiority over 24-bit - both are considered "Full precision"...

Epic Games – Tim Sweeny
Only a marketing guy would consider 24-bit floating point to be "full precision"! 32-bit floating point has been an industry standard for 18 years, since the original IEEE 754 specification was ratified. Besides the obvious need for 32-bit precision in any sort of high-range calculation (such as position vector math, world transformations, etc - all things one is likely to do frequently per-pixel), we feel there's a huge advantage to using industry standard arithmetic in its predictability and reproducibility across different GPUs. The 24-bit precision mode allowed by the DirectX9 spec was backed by a reasonable cost/performance/precision judgment back in the early days of DirectX in 2002-2003. But nowadays there isn't a good reason for hardware to support less than full 32-bit floating point precision.


----------------------------------------------------------------


GD: A lot of attention has been paid to NVIDIA's support of Pixel Shader 3.0 - can you specifically think of anything PS 3.0 can be used for that can't be done in PS 2.0?

Andrey Khonich – Crytek, Makers of Far Cry
In VS3.0 Shader model actually is possible to support general displacement mapping (with smart Shader design when vertex Shader has to do something during waiting for texture access).

In PS3.0 shaders it’s possible to decrease number of shaders using dynamic branching (one Shader for general lighting) and in such way decrease number of Shader switches and as result increase speed, and also we can utilize dynamic conditional early reject for some cases in both PS and VS and this also will increase speed. As to NV40 generally possible to use co-issues better to take advantage of super-scalar architecture (we can execute 4 instructions per cycle in a single pipeline).

- Handle several light sources in single pixel shaders by using dynamic loops in PS3.0.
- Decrease number of passes for 2-sided lighting using additional face register in PS3.0.
- Use geometry instancing to decrease number of draw-calls (remove CPU limitations as much as possible).
- Unrestricted dependent texture read capabilities to produce more advanced post-processing effects and other in-game complex particles/surfaces effects (like water).
- Full swizzle support in PS3.0 to make better instructions co-issue and as result speedup performance.
- Increase quality of lighting calculations using 32 bit precision in pixel shaders on NV40.
- Take advantage of using 10 texture interpolators in PS3.0 Shader model to reduce number of passes in some cases.
- Easily do multiple pass high-dynamic-range rendering on FP16 targets.
- Speed up post-processing a lot by using MRT capabilities and mip-mapping support on non-power-of-two textures.





well.. it was very interesting the reading.. :)
i almost imagined TIm slapping that guy face.. by the way he asked that question :D since he have been yelling for FP32 to be supported by all IHV for a quite some time now.. he needs it for his game. :)

reading the word Quake4 in that interview make me Drool.:drooling:
and even without ever seeing a single in game screenshot.. :D
there are also other developers that downplay PS3.0 ...Funny how diferent game developers can have opposite views about the same question.. :)

in the end ,i think that they are all Right.. SM3.0 will be a huge improvement over SM2.0 for the gamedevelopers that will take advantage of it ,in their games. For the developers that will not use it ,in ther games ,it will not make any diference at all. :)

[Brick_Top]

Hum... Even if It's true it sounds like marketing nv40 campaign, just like the one ATI must be starting to downplay SM3.0.

[Nv40]

Hum... Even if It's true it sounds like marketing nv40 campaign, just like the one ATI must be starting to downplay SM3.0.

yes.. but marketting doesnt necessarily means ,that its false.. isnt it?
Ati have been marketing PS2.0 for along time.. and now Nvidia with PS3.0. Nvidia have been maretting Doom3 as the next quantum leap in graphics.. ATi have been doing it with Hl2.. so not everything that is marketed by a company is necessarily a fake. ;)

[DSC]

Unreal Engine 3.0 makes HL2 look like obsolete stuff.....

:lol2:

[nForceMan]

http://www.gamers-depot.com/interviews/dx9b/001.htm


Nice find NV40! :thumbsup:

[oqvist]

Of course movie effects are done with 32-bit precision but those don´t have to take into account performance hit´s or such since that isn´t done in real time.

For me the question of weither the Shader Model 3 will be of much use is if the hardware that is running it is fast enough. There is no way games optimized for Shader Model 3 will run faster than games running in PS 2.0 mode.

And it´s a bit worrying that ATI says that their PS 3.0 really hurts their performance when everything seem to point up that performance wise the X800 PRO will be about as fast as the 6800 ULTRA...

In the end it´s yet to see how much it means. It´s natural nVidia want to market it as a very important feature because of ATI decided to take it out of their design and ATI will of course do exactly the opposite.

But the good thing is that this time we may be able to see the results much quicker than we did with the PS 2.0 support which not until the last 6 months or so has really delivered.

[Daneel Olivaw]

Is it me or do people keep assuming FP16 means 16 bit, when it means:
FP16: 64bit
FP24: 96bit
FP32: 128bit

I have an OpenGL programmer friend that said:
"Cool man, imagine, un Z-buffer en 128 bits! Je sais même pas quoi faire avec un Z-Buffer en 64 bits, je suis content avec 32 bits!"

[reever2]

Unreal Engine 3.0 makes HL2 look like obsolete stuff.....

:lol2:

Especially since the *engine* is being released in 2005/2006

[oqvist]

Yeah kind of pointless using it in the marketing when the NV40/R420 is well below the recommended specs in the first place.

[Templar]

I bet that ATi are going to get their money back from Vivendi, Half Life 2 is not shipping this year HEHHE.. not that I care any more.. plenty of super cool games coming out and out already.. FARCRY..

Simon

[svetli_pp]

First I have to say NVidia won this time imho. Like a developer said SM3.0 is something that removes programming barriers, which does not mean that it will be fully utilized at least until even next gen card several times faster than NV40 hits the market. And its for the barriers of performance. Longer shader code will require faster GPUs to achieve decent fps. Though I'm sure we will see some mild benefits from SM3.0 in the forthcoming titles like Doom3, Farcry etc. I don't expect to see much better graphics in SM3 if any at all. ATI do not lose because of lack of functionality but because of the desire of hardware enthusiasts to be on the peak of features when they buy new stuff. Its similar to AGP4X/8X thingie. For current games at playable resolutions 4X is enough but who will buy MB with 4X support only. We'll see great fuss in the near future over this topic, and I suppose things might clear a bit.

[maxbero]

Especially since the *engine* is being released in 2005/2006

early than hl2? :)

[MontoyaSG]

First I have to say NVidia won this time imho. Like a developer said SM3.0 is something that removes programming barriers, which does not mean that it will be fully utilized at least until even next gen card several times faster than NV40 hits the market. And its for the barriers of performance. Longer shader code will require faster GPUs to achieve decent fps. Though I'm sure we will see some mild benefits from SM3.0 in the forthcoming titles like Doom3, Farcry etc. I don't expect to see much better graphics in SM3 if any at all. ATI do not lose because of lack of functionality but because of the desire of hardware enthusiasts to be on the peak of features when they buy new stuff. Its similar to AGP4X/8X thingie. For current games at playable resolutions 4X is enough but who will buy MB with 4X support only. We'll see great fuss in the near future over this topic, and I suppose things might clear a bit.

if games based off SM3 runs better due to efficient code use the tables will be turned

[nForceMan]

early than hl2? :)
:nutkick:
:rofl

[Jethro]

Wow this could have more of an impact than i woulda thought..

These programming guys really like the prospects and it looks like the only way for them to explore and tweak ATM is on Nvidia hardware.

Reguardless of how far out these features are Nvidia is getting a serious foothold for the "way its meant to be played" by supporting such high level tech that excites the dev and the gamer. Nobody knows for sure but it seems like this could really hurt ATI this go round no?

[Jethro]

man if this isnt saying what many of us have been TRYING to say...

Mark: I think today it's more of a personal preference, seeing as only a handful of games currently support either PS2.0 or PS3.0. But by the end of this year there will be enough games using PS2.0-PS3.0 for people to decide which card better meets their needs. I would personally go for the card with a higher feature set, because then it will have a longer life span. Besides, if card X runs Q3 at 170fps and card Y runs Q3 at 180fps, who cares? As long as I get at least 60fps, I'd rather have prettier graphics.

[Nv40]

Wow this could have more of an impact than i woulda thought..

These programming guys really like the prospects and it looks like the only way for them to explore and tweak ATM is on Nvidia hardware.
Reguardless of how far out these features are Nvidia is getting a serious foothold for the "way its meant to be played" by supporting such high level tech that excites the dev and the gamer. Nobody knows for sure but it seems like this could really hurt ATI this go round no?

absolutely... and they know it. i think ATi is more worried more at this time about Sm3.0 than by the performance of the Nv40 . :D
because in Sm3.0 not only Nvidia can get very quickly most Ps2.0 games to be patched for 3.0 ,with a few little gimmics (at least) ,but also they can get huge improvements in performance over Sm2.0 path ,if it is well implemented. here is another developer.. opinions..


In my opinion, Shader Model 3.0 is a huge step forward compared with Shader Model 2.0. Shader Model 3.0 adds dynamic branching in the pixel shader and while it's not required, I'm expecting the major IHVs to provide a complete orthogonal solution (FP16 texture filtering/blending) in their next HW iteration.

At this point (things may change), I'm expecting that Splinter Cell - X will only support SM 1.1 and SM 3.0 when it comes out.

-Very significant performance improvement because of dynamic branching in the PS unit.
-Orthogonal FP16 operations
-Market (can't discuss that yet)

SM 3.0 is going to be good enough for some time. There is only one big step left (before GPU start evolving just like CPUs --> performance only) that should allow classic global illumination algorithms to be efficient on GPUs. I doubt SM 4.0 will provide that.
_________________
Dany Lepage
Lead Programmer
Splinter Cell - X
UbiSoft Montreal




VAlve ,Idsoftware ,Stalker developers,Bungie are a few others i will like to hear..

[jbirney]

Notice how developers on the TWIMTBP program raved about PS3.0. Where developers that were not did not seem to care? I really liked Barts one commit:

Bart: We like to ensure our games work on the majority of video hardware regardless. The card vendor certification initiatives seem to affect marketing more than development.

Its about time some one said this.

[jAkUp]

I bet that ATi are going to get their money back from Vivendi, Half Life 2 is not shipping this year HEHHE.. not that I care any more.. plenty of super cool games coming out and out already.. FARCRY..

Simon

Its coming this summer, I would put money on it.

hmmmm

Currently it appears the impact of 3.0 on developers will be minimal. Therefore it's unlikely the consumer will see significant benefits from 3.0 for quite some time, especially since the improvement of 3.0 over 2.0 isn't as great as 2.0 over 1.0.

[eesa]

okay, keep holding your breath for those hyped PS3.0 speed improvements... Of course nothing is for sure, but if I knew any of you personally right now I'd put down money that says it's just a checkbox feature. Sure I could lose but I think the odds are in my favour :)

[Jethro]

I knew someone would extract some of them comments made and hold em up and as if to say see!!, i shouldve known who would bite first lol.

any intelligent person can obviously see from that group of Q and A's there is no denying that 3.0 is exciting enough to some Devs to make it something to take seriously. the obvious fact here is 3.0 will be in development now and optimised for Nvidia hardware well before ATI even supports it. <<<this could be a big deal in this round! im not saying that its important right this minute but this could give Nvidia pole position again over the next few years.

but that would require one seeing the big picture and not focusing on the little view.

[eesa]

any intelligent person can obviously see from that group of Q and A's there is no denying that 3.0 is exciting enough to some Devs to make it something to take seriously. the obvious fact here is 3.0 will be in development now and optimised for Nvidia hardware well before ATI even supports it. <<<this could be a big deal in this round! im not saying that its important right this minute but this could give Nvidia pole position again over the next few years.

but that would require one seeing the big picture and not focusing on the little view.

that would be absolutely true, if there was actually something to 3.0
If 3.0 were some revolutionary feature, then nv is at a huge advantage as everyone is using their hardware to develop the next generation of games. However, that's a very big, gigantic if. At little too big if you ask me.

[Nv40]

okay, keep holding your breath for those hyped PS3.0 speed improvements... Of course nothing is for sure, but if I knew any of you personally right now I'd put down money that says it's just a checkbox feature. Sure I could lose but I think the odds are in my favour :)

i dont know why i have the feeling the NV40 will end being faster ,that the R420xt and with the SM3.0 features.. but i could be wrong :D

but this strange behaviour of ATI , lately.. downplaying the NV40 , claiming how faster they will be than Nvida ,they have been doing more claims about how they are going to own Nvidia.. while Nvidia remain quiet. i will not be surprised if Nvidia have not shown yet their full hand..but only a finger.. hehe . :). i think that there will be many people dissapointed if raise their expectations too much about ATI.

[Jethro]

that would be absolutely true, if there was actually something to 3.0 If 3.0 were some revolutionary feature, then nv is at a huge advantage as everyone is using their hardware to develop the next generation of games. However, that's a very big, gigantic if. At little too big if you ask me.



tru its a bit early to make such calls ill give u that. Kinda ironic tho how ps 2.0 was downplayed with much less enthusiasm than this and look how that resulted. Now we have ATI scoffing at Nvidias poor PS 2.0 perf. The irony..

Nvidia hasnt exactly lost ther azz either peeps, they got one upped in the 5800 vs r300 battle, and gave up some ground since. which is not a bad thing for us by no means however the 5900 was not a flop, in fact sales have been pretty darn good lately and lets not forget the tons of other Nvidia hardware on the market.

The 6800 to me represents a well thought out response to an underestimated foe in the graphics arena. And in my view in the next coming days and weeks ahead your gonna see this thing steamroll into a huge success for Nvidia. The current buzzes are only the beggining.

Does this not remind folks of the r300's entrance into the market in some ways? where we all waited for Nvidia's response, dx9, ps 2.0 talk etc.

[Socos]

Unreal Engine 3.0 makes HL2 look like obsolete stuff.....

:lol2:

True, but Unreal 3.0 won't be out until 2006. I am sure there will be plenty of wonderful engines out by that time..