Kars, i think you are missing out what albionus tried to say.

In fact, both of you re correct, but are not dealing upon the same subject.

- On a technical side (your side), what will be do-able on PS3 regarding physics and graphics, given its distributed  parallel computing model is an order of magnitude below what will be acheivable on Wii.

-  On another side (business?), nothing is unmanageable if your reduce assets enough to meet the hardware limits.

One thing people don't realize is that you can span quite a massive performance difference with good design practices ...

For years (on the PC) developers have had to support very low minimum requirements while still providing the impressive graphics to sell the game to people who own bleeding edge PCs. The general strategy is to produce 3 levels of detail for most of your assets (this sounds like a lot more work then it is mainly because there are a lot of tools for downgrading graphical assets).

So on the Wii you would have a graphical asset model and textures while rendering the game at 480p with 2xAA; on the PS3/XBox 360 you would have a similar model that had 2 to 4 times as many polygons, 2 to 4 times the texture detail, high level pixel shader effects, while rendering the game at 1080i/720p with 8xAA and 16xAF. The XBox 360 and PS3 game would look far better but the core gameplay would be identical on all systems.

Why do you not see MySims as a AAA title?

Do you see Guitarr Hero as a AAA title? Because I do and Guitar Hero 3 is also coming to Wii.

It might be easier to discuss this if you could explain what a triple A title for you is.

"So on the Wii you would have a graphical asset model and textures while rendering the game at 480p with 2xAA; on the PS3/XBox 360 you would have a similar model that had 2 to 4 times as many polygons, 2 to 4 times the texture detail, high level pixel shader effects, while rendering the game at 1080i/720p with 8xAA and 16xAF. The XBox 360 and PS3 game would look far better but the core gameplay would be identical on all systems."

That fact is i think the majority of developers will be "lazy" enough for things happening this way.

But effects and calculations will be doable on Xbox360 (parallel computing capable) and even better on PS3 (distributed parallel computing capable, as said by Kars, in which he is perfectly correct) that you won't be able to do on Wii.

Will that change the paradygm on presentation for the games ? If i had to give an answer now, my answer would be NO. Because, IMO, the next shift will be when raytracing will be acheivable real-time. Raytracing will not be possible before next-gen (PS4 and Xbox720), and even at this moment it is actually questionnable ...

edit : changed "raycasting" to "raytracing" after Harry's comment ...

You mean raytracing ...
Raycasting was what John Carmack used to make Doom ...

I just don't get all this stuff about the Cell and AI. Artificial intelligence is all about decision making, choices - in a word: branching. The Cell excels at physics, numerical simulation - not at branching. The amount and quality of the information based on which choices can be made improves, that's for sure. The decision process however, I don't think so. Having fast a dedicated core for AI, simulating hundreds/thousands of actors' decision processes while evolving the world, as is possible on the 360, is great. Separating these concerns (the world model, its views and controlers) the way it can be done on the 360, is a studied problem. In the PS3 don't see any number of SPEs being able to handle AI actors independently of the PPE. Obviously, if most of the time is spent with physics and numerical stuff, being able to offload that to the SPEs, gives you more free PPE time to do AI. But that does not make the Cell good for AI - it just means that you don't do that much AI and the Cell is good for everything else you do.

Well, theren are cases where you can see a number of statements as vectors ande manipulations as Vector Additions and Multiplications. This is something that a SPU can handle quite well, but this kind of code would be very hard to read and debug andf you alwys have to translate into these vectors and out of them, what normally reduces the usability considerably.

But even the PPE itself or the Xbox 360 isn't really suited for branching intensive code. The pipeline is quite deep and its brnaching prediction is rather simple, which makes it slower than many other processors for branching intensive code.  

Woot, another member who understands silicon. :) I'll probably participate more in this thread after I've had some hands-on Cell programming experience (which should begin later this week). But I would like to say a few things about the development process for using SPEs. I think if you're going to write code for a platform with 8 processors, you should design your engine with parallelism in mind. Writing everything for one processor and then parallelizing it later for performance is a valid approach, and will work very well for cross-platform games. But you're not going to get the same level of performance that you would if you think in terms of parallelization from the beginning. Writing parallel code for the Cell is much harder than just starting another thread on a shared-memory SMP system, since each SPE has its own local store that must be managed separately. There may be design decisions you can make before you write a single line of code that have profound effects on the kind of performance you can get out of the Cell versus a single-CPU machine or a shared-memory machine.

I disagree ...

The general guideline of all development is that 5% of your code will be running 95% of the time; the vast majority of your code will be run very rarely and does not really need to be optimized at all (that is, unless it is running at an unreasonable rate).

I've not worked with the Cell but I have worked with DSPs and you're going to get most of the performance benefits from the Cell (that can be achieved in a cost effective way) by developing your engine in ANSI C++, creating Assembly Language versions of most of your math libraries, and then parallelizing small portions of frequently called code and running it on a SPE.

Regardless of which approach you take you will find that a major problem will be preventing the SPEs from being idle ... Even if you optimize most of your code to take advantage of the SPEs most of the time you will only be using one or two of the SPEs, and when you are using most (or all) of the SPEs one will likely be overloaded while another is starved for work.

Nintendo's got a lot of positive buzz, but I'm not seeing where it's all of a sudden become hip. Even if we were to assume that you could easily port something from 360/PS3 to Wii, developers don't appear keen to test it. The Wii version of any three platform game has been designed on the Wii, generally as a spinoff and/or by a different team. MySims isn't about pushing the boundaries. It's about finding another potential Sims market to exploit. In terms of sales, I would expect them to be modest at best. Guitar Hero's probably fairly cheap to produce, though it will sell a lot.