HI-JACKED!!!

@MikeB
really man, you have to starrt your own threads rather than hi-jacking others...

MikeB said: @ Entroper The topic was about PS3s being used to crack passwords Which is achieved by the calculating power of the SPEs. On a gaming orientated website like this I don't think it's bad to address its relevance to gaming.

You're not addressing anything.  You're spamming.  There's a difference.  You want to engage people in discussion, I have no problem with that, but I have yet to see that actually happen.

NJ5 said: Shit MikeB, did you really have to fetch your txt file where you have the propaganda that you copy paste to all threads even remotely related to the Cell?.

QFT and LOL

NJ5 said: Yes, the Cell can be made to run games well, but at what cost in terms of development? The thing which Microsoft brought to the industry was knowledge and willingness to help developers be more efficient, Sony should learn that from them. No matter what some exclusive developers say, asymmetrical processors with glorified DSPs running on low memory must be a bitch to program some algorithms with. The PS2 was a difficult architecture too, but developers had to live with it because it was so successful. Sony is probably hoping that the same happens with the PS3, but it's not guaranteed.

There's definitely something to be said for staying close to familiar architecture, but sometimes you just have to bite the bullet and go with different, superior architecture. I'm not a programmer but from what I understand most of the complaints about PS3 vs. 360 architecture stem from the fact that the 360 has a very PC-like processor, while the PS3's Cell is an entirely new and rather complicated beast. It's true that it has taken a lot of effort for developers to get truly superior performance out of the PS3, and those efforts have been expensive. But once the groundwork has been laid, it should become easier to get that performance out of the PS3, right? If the PS3 is significantly more costly to develop for than the 360 (I don't know), it may well be true that that will be the case for the entire lifespan of both consoles. However, I think that over time the PS3 will create and widen a performance advantage as developers acclimate to the hardware. HAVING SAID THAT MikeB, not every single thread relating to the technical abilities of the PS3 needs to turn into god damned trench warfare between you and whoever takes offense to your propaganda. Please keep that in mind next time you get a twitchy trigger finger on "Ctrl", "C", and "V".

Hey guys, do you really not have anything of interest to say?

AFAIK, the original poster hasn't complained to me regarding my input. If he did I would have stayed out of this thread. Some people may be a little upset about not owning a PS3 or that their XBox 360 or Wii doesn't have a Cell processor.

The Cell is quite remarkable (all these achievements are really quite amazing!) and I think this surely warrants some more in depth discusion. I think I provided quite a nice little summary in this thread (yes, collecting stuff I have been saying before), but again the same old fools want to drag this thread into something personal, not even remotely relating to the calculating power of the Cell. Keep up the good work guys!

@ Final-Fan

One problem is the general design of legacy game engines (not designed for parallel processing, which is achieved by multi-threading, Amiga and especially BeOS later pioneered in this area), especially if they use too much of Microsoft's closed technologies. Porting games from DirectX/Win32 over to other platforms has always been a hurdle for developers, it can be done like some developers I know well have ported high profile commercial Win32/DirectX based games over to Mac, Linux and AmigaOS, but this takes a lot more effort and they rewrote and adapted some of their middleware/3D driver systems to simplify the process.

The Cell is very easy to develop for if you don't use the SPEs at all, but like with almost everything which is more advanced (such as the introduction of multitasking operation systems, you shouldn't assume your program is the only one running and you may want to allow interaction with these programs) when the system becomes smarter and more feature rich there are usually more things to take into consideration when developing software. Having said that many PS3 developers are eager to learn and love to explore what the Cell can achieve and they are.

The Cell and the PS3 in general isn't really that hard to develop for, in fact all developers I know of said it's actually quite a bit easier to develop for than was the case with regard to the PS2. If you know anything about programming I recommend to look at IBM's website for some nice Linux related tutorials.

MikeB wrote:

From a recent interview regarding Burnout Paradise:

"We've always loved Sony machines and we've loved PS3 from the start, we think it's a fantastic machine. We felt very strongly about that.

So from the beginning we were always going to lead on PS3 and our transition between the two machines has been really seamless and straight-forward, where as other people have lead on 360 and found it hard to get on PS3."

We have heard many similar statements from other multiplatform developers, including from the devs behind the hot selling multi-platform game Call of Duty 4.

Pretty much says what I said above: 

"For that, you can include Xbox 360 and high-end PC games, and exclude PS2 and Wii. However, since PS3’s Cell processor allows MORE features - better physics, more complex graphical processing, lighting or sound, etc. — there is inevitably going to be more cost in supporting those extra features. It’s not that PS3 is harder to write for, it’s just that you can do more with it." 

http://blog.us.playstation.com/2007/06/13/is-the-ps3-really-harder-to-develop-for/ 

 

GRAW developer:

"It's wrong to say it's harder to code on the PlayStation 3, it's just something that needs to be learnt," 

 "Developers might says [sic] it's harder because it just takes time to understand the technology, we're still early in the lifecycle." 

Final-Fan said: NJ5 said: Yes, the Cell can be made to run games well, but at what cost in terms of development? The thing which Microsoft brought to the industry was knowledge and willingness to help developers be more efficient, Sony should learn that from them. No matter what some exclusive developers say, asymmetrical processors with glorified DSPs running on low memory must be a bitch to program some algorithms with. The PS2 was a difficult architecture too, but developers had to live with it because it was so successful. Sony is probably hoping that the same happens with the PS3, but it's not guaranteed.   There's definitely something to be said for staying close to familiar architecture, but sometimes you just have to bite the bullet and go with different, superior architecture. I'm not a programmer but from what I understand most of the complaints about PS3 vs. 360 architecture stem from the fact that the 360 has a very PC-like processor, while the PS3's Cell is an entirely new and rather complicated beast. It's true that it has taken a lot of effort for developers to get truly superior performance out of the PS3, and those efforts have been expensive. But once the groundwork has been laid, it should become easier to get that performance out of the PS3, right? If the PS3 is significantly more costly to develop for than the 360 (I don't know), it may well be true that that will be the case for the entire lifespan of both consoles. However, I think that over time the PS3 will create and widen a performance advantage as developers acclimate to the hardware. HAVING SAID THAT MikeB, not every single thread relating to the technical abilities of the PS3 needs to turn into god damned trench warfare between you and whoever takes offense to your propaganda. Please keep that in mind next time you get a twitchy trigger finger on "Ctrl", "C", and "V".

I actually agree with FinalFan on this one.  I am a programmer, and your understanding is basically correct.  A symmetrical design is much more familiar to developers, and an asymmetrical design presents a lot of new challenges, mainly in thread control (not so difficult) and memory management (more difficult).  The fact that the SPEs cannot directly address main memory is simultaneously the Cell's greatest strength and its greatest weakness.  The fact that you have to initiate DMA transfers any time you want to read or write to main memory makes the coding more difficult, but the fact that you can see where all the transfers are and carefully analyze the flow of data makes optimization somewhat automatic.  Somewhat, because you can still code it poorly, but if you're good enough to code it at all, you're probably good enough to do it well.

 

@MikeB... it's not even that I disagree with you on any particular point.  It's that you seem to have made it your mission in life to loudly proclaim the greatness of the Cell from every mountaintop and disprove every slightly negative thing anyone says about it.  Some of us don't have anything to prove, we just want to have a conversation.  I understand that there are a lot of myths, misunderstandings, and FUD regarding the Cell, but this is true of every important piece of technology.  In fact, if you go back a few months on these boards, you may find that I've argued against a lot of the FUD, but you have to know how to distinguish between FUD and reality, and when to draw the line between dispelling myths and being the crazy poster boy.  I'd actually be on your side on some of these issues if you weren't so annoyingly fanatical.

@ Entroper

Yes, more difficult for low level game engine and middleware developers, but those aren't actually the ones complaining, it's mostly those who are dependent on 3rd party middleware solution providers who haven't yet finished adapting their code.

From a performance perspective an asymmetrical design offers great benefits,

Insomniac's igPhysics:

Benefits of SPU Shaders to igPhysics
● Pipeline well defined and completely SPU-driven
● SPU processing completely asynchronous
● Data well-organized and well-defined.
● No (or minimal) PPU intervention



No like I said way back in 2005, I fully well realize the hurdles developers have to go through. But I care more about the potential end results and I believe within a couple of years most of the guys here will do so too.