Ati and Nintendo GC successor GPU

[MfA]

Considering Moore's history at IBM and it being in total contradiction with the Forth Chuck Moore's resume it seems rather unlikely.

A search on ibm.com places him as an employee in the IBM enterprises system group in '99. Funnily enough he is also referenced as both Charles and Chuck (the Forth Chuck Moore was originally also Charles). Thankfully in reference to the same project with Vikas Agarwal ... so we can be reasonably sure there arent 3 C. Moore's in question The research he is related to is related to processor architecture, so we can be reasonably sure he and not the Forth Chuck Moore is our man.

I like stack machines, but they are very hard to compile to ... since they have never been very popular the research into compilation techniques isnt exactly widespread either. Chuck Moore is also rather religious about how he wants his stack architecture, and his preferred one (0-operand, no hardware support for stack overflow) to my eyes is the worst possible one for compiled code (I think 1-operand is better).

 

[DeanoC]

ok i see... still how is a PowerPc970 going to compare to Cell? PS3 will most certainly have a total of 1Tflops spread between CPU and GPU as we've seen...

the thing is, as big and good as they are, the major investment put into Cell is enough for me to censider it the "main" project IBM will have in the future, especially considering that it will be included in EVERY Sony appliance released, not only PS3...

If your prediction is right, that Cell has 1 TFlops for CPU and GPU, then Cell may actually be appear quite SLOW compared to a PowerPC970 or Intel based competitor, because just like the PS2 where VU0 and VU1 are given specialised jobs (usually physics and graphics) and the main CPU is actually quite slow, Cell may end up being very similar, great for the vector processing operations (which only really covers physics and graphcs) but lacking at general logic processing (i.e. executing 'standard' C++ game code and control logic) which is often the limiting factor in real games (outside benchmarks).

So if Xbox2 or GCN2 is able to ship a powerful GPU (which are essentially vector processing units) which can be used for both graphics AND physics, it would leave there main CPU for the job of running the game. If they are able to get a couple of GHz with large cache's then it may have considerable more 'apparent' CPU power than Cell.

Its been said many times, but its so true. IF Cell (assumming the patent is about Cell etc) is to succeed in its hyped promise, then Sony are going to have to build an awesome development enviroment and real-time OS, to manage and ease the whole multi-threaded architecture. Personally if I was Sony, I think seriously about purchasing somebody like QNX (IMHO QNX is far more suited to the stable real-time enviroment that console needs than linux, but then linux is cheaper) to provide the kernel primitives and Code-warriors to built the IDE/compilers.

If Sony don't provide the mother of debuggers, vectorizing compilers and thread primitives, all those extra CPU's are going to be wasted until the later generations. Thats where a slower but more general system (like Intel or PowerPC) could win.

Its even in IBM 'interest' to drop Nintendo a few 'hints' on how well Cell is going. Bigger processors mean more money going to IBM.

IBM is a win/win situation for the next generation unless MS wins....

 

[Vince]

The research he is related to is related to processor architecture, so we can be reasonably sure he and not the Forth Chuck Moore is our man.

Ok, then I stand corrected - all five of them are IBM employees, which makes sence. Thanks Marco.

Chuck Moore (Forth) is just one of those people that once you hear of you don't forget and 25x was inb my mind especially. "Religious" is almost an understantment with his insane desire to push Forth. So Chuck, what's the most advanced language? Forth. Chuck, I haven't looked at Forth in like 8 years, can you port C-based programs to Forth easily or automatically? What is this C you speak of? Forth is easy and the last thing you'll need to learn. Umm.. ok buddy...

 

[cybamerc]

Paul:

> ... despite dozens of currrent articles and Sony stating it would be
> inside the system?

I love how you Sony fanboys cling on to every report stating that Cell will be in PS3 while continously ignoring the numerous reports stating that it won't or may not.

I'm just not convinced that Sony will manage to get everything ready in time. The hardware will obviously be finished soon but that is the easy part. Now they have to figure out how to program it and optimize production so they can pump the chips out by the millions while making it affordable.

 

[notAFanB]

Paul:

I love how you Sony fanboys cling on to every report stating that Cell will be in PS3 while continously ignoring the numerous reports stating that it won't or may not.

correct me if I'm wrong here but there are a few reports which state that CELL 'may not' be ready for PS3 (or that it may be delayed).

there are no such reports which categorically deny that CELL was intended/will be for the PS3 platform thus borderring on sceptism is understandable, outright denial that sony can make the deadline is a bit much?

 

[Grall]

I love how you Sony fanboys cling on to every report stating that Cell will be in PS3 while continously ignoring the numerous reports stating that it won't or may not.

Personally, I think it was totally unneccessary to employ the F-word in this situation, it was absolutely completely uncalled for. If I was a mod, I'd slap your ass for doing so.

Now, on to your post... I think it's pretty clear Cell WILL be in PS3 for many reasons.

Just for starters, we KNOW that Cell is on-track. There's no delays reported, quite the opposite. Then there's the matter of what ELSE could Sony possibly use Cell for? Why would they pour billions into the development of an extremely microprocessor destined to be completed around the time of PS3's launch and then NOT use it in PS3, the only product of theirs that could possibly find a use for it? That just sounds totally stupid to me and to anyone with half a brain (but maybe not to you?). Lastly, if Cell does not end up in PS3, what else could they possibly use to replace it? They have presented no alternatives, and in fact there are no alternatives anywhere near in Cell's league. Switching out Cell late in the game would force throwing away all the work they'd done on the system so far. They're not idiots over there at Sony, they wouldn't do that.

I'm just not convinced that Sony will manage to get everything ready in time.

"I love how you non-Sony fanboys cling on to every report stating that Cell won't be in PS3 while continously ignoring the numerous reports stating that it will."

You enjoy throwing stones in glass houses man?

*G*

 

[Josiah]

So if Xbox2 or GCN2 is able to ship a powerful GPU (which are essentially vector processing units) which can be used for both graphics AND physics, it would leave there main CPU for the job of running the game. If they are able to get a couple of GHz with large cache's then it may have considerable more 'apparent' CPU power than Cell.

I'm not so sure about this anymore. MS had a very specific goal to outclass PS2 with XBox, and they ended up with this very expensive piece of hardware thats still being sold at a loss today. Maybe with XB2 they will be willing to settle for simply "comparable" graphics instead of "better" graphics, on hardware that is cheap for them to manufacture. In that case they will lose their one advantage (better hardware), but I just don't see them making another system thats uber-expensive to manufacture.

Its been said many times, but its so true. IF Cell (assumming the patent is about Cell etc) is to succeed in its hyped promise, then Sony are going to have to build an awesome development enviroment and real-time OS, to manage and ease the whole multi-threaded architecture. Personally if I was Sony, I think seriously about purchasing somebody like QNX (IMHO QNX is far more suited to the stable real-time enviroment that console needs than linux, but then linux is cheaper) to provide the kernel primitives and Code-warriors to built the IDE/compilers.

If Sony don't provide the mother of debuggers, vectorizing compilers and thread primitives, all those extra CPU's are going to be wasted until the later generations. Thats where a slower but more general system (like Intel or PowerPC) could win.

You would *THINK* that they would have learned this lesson during the PS2s growing pains. Simply because they went through that experience, with all the complaints from developers, and the lackluster titles, I think they'll be better prepared with PS3.

 

[Paul]

I love how you Sony fanboys cling on to every report stating that Cell will be in PS3 while continously ignoring the numerous reports stating that it won't or may not.

I love how Xbox fanboys continue to cling to a Designchain article which has been proven moot so many times it's not even funny.

Numerous reports? Do you mean spong, the inquirer, and teamxbox? Which actually took the designchain article and put their own spin on it?

Funny, Kutaragi just stated Cell is ON TIME and they are going to have a prototype by march 2004.

So we should take Mr. Spong's words over the MAN in charge of Cell? Where is the logic in that, it must be hidden behind that huge mound of BS outback.

Or should we take Sony's own words, dozens of articles which were made after the Designchain article that state that Cell is indeed the processor that will power PS3.

I'm just not convinced that Sony will manage to get everything ready in time.

Yea and let's ignore the fact that a Cell prototype is right upon us...


Let's see your evidence of Cell not being in PS3, and than I will state mine. We will see who walks away embarassed.