Predict what processor will be used for the Xbox 360

Status
Not open for further replies.
True - Toshiba, and Sony as well down the line, both have plans for Cell-equipped televisions. What else the chip might end up in that's pure consumer electronics, we'll just have to see. If I remember correctly, Sony's big Cell push will supposedly be in 2007, when the PS3 launch is safely behind them.
 
Sorry if this has been asked before but i'll ask anyway. Could MS, IBM & ATI engineer the X2 to give 1Tflop of processing power. I'm not talkin about just a powerfull cpu or gpu, more like an overall system design that could or would get u too 1Tflop of computational power. I would like to get your ideas as to how it would be possible, what kind of hardware and software optimisaition you'd need.
 
Shifty Geezer said:
As far as I can recall, the 1 teraflop figure was for Cell's capabilities or something. It wasn't specifically applied to PS3 by KK, but the media started pinning it as a benchmark figure for the system. No-one here's expecting 1 TFlop in PS3, at least not conventional processing.

It seems to me that MS's 1 Teraflop figure was a knee-jerk marketting rection to Sony's. Any number you can use, we can use better! I don't think the numbers are at all relevant as they neither offer fair comparisons of the systems (because both parties will be making up whatever numbers they can to out-value their opponents, including even the system's number - 360>3!!) nor provide useable benchmarks that give an indication of what the systems' realworld performances will be like. Kinda like this generation, where fans toot numbers showing one system's better than the other, but I can't see much difference between them all myself.

The 1TFLOPS figure for CELL was based on a CELL architecture implementing 4 PPEs and 32 SPEs... which was what KK was dreaming for PS3. Unfortunately for him it won't be possible until around 2010 according to IBM. They will probably need 45nm to 32nm fab process to achieve that in a single mass produced chip with good yields.
 
vblh said:
Sorry if this has been asked before but i'll ask anyway. Could MS, IBM & ATI engineer the X2 to give 1Tflop of processing power. I'm not talkin about just a powerfull cpu or gpu, more like an overall system design that could or would get u too 1Tflop of computational power. I would like to get your ideas as to how it would be possible, what kind of hardware and software optimisaition you'd need.

For the last time - GPUs lend themselves very well to ultra high (inflated?) flop ratings. The R500 alone will likely be rated in ~1Tflop territory by ATi and MS. I mean, some were classing NVidia's latest graphics card as a 1Tflop chip! When you understand that, it's not hard to see how MS and their partners could magic up a "over 1Tflop" figure..
 
It's not hardware and software optimisatins you need to reach 1 teraflops - it's creative statistics and innovative number extraction routines!
 
Titanio said:
vblh said:
Sorry if this has been asked before but i'll ask anyway. Could MS, IBM & ATI engineer the X2 to give 1Tflop of processing power. I'm not talkin about just a powerfull cpu or gpu, more like an overall system design that could or would get u too 1Tflop of computational power. I would like to get your ideas as to how it would be possible, what kind of hardware and software optimisaition you'd need.

For the last time - GPUs lend themselves very well to ultra high (inflated?) flop ratings. The R500 alone will likely be rated in ~1Tflop territory by ATi and MS. I mean, some were classing NVidia's latest graphics card as a 1Tflop chip! When you understand that, it's not hard to see how MS and their partners could magic up a "over 1Tflop" figure..

The problem with your post is not that you may not be right... its that your cynical and critical of MS and ATI when you should be across the board.. and yes you criticised NV also but there is no way in hell that Sony and NV will not say PS3 is a 1 teraflop+ system also (by whatever mix of parts they end up with).

And if the 1 teraflop number is true then what? :?:
 
Titanio was replying to post asking explicitly about MS, IBM and ATI manufacturing a 1TFP XB360, and gave a suitably explicit reply. Can't see anything wrong with that. It's not as though either was claiming Sony/NVidia weren't guilty of the same; just that that wasn't the topic of conversation
 
Shifty Geezer said:
Titanio was replying to post asking explicitly about MS, IBM and ATI manufacturing a 1TFP XB360, and gave a suitably explicit reply. Can't see anything wrong with that. It's not as though either was claiming Sony/NVidia weren't guilty of the same; just that that wasn't the topic of conversation

Well then he didnt answer the question. He was just explicit about the graphics card particularly the R500 and how MS and their partners "magic" up a number (which is a funny ass term BTW :p )... I just think that everyone "magics up" the best number possible is all...

Dont worry we will go through this same conversation again in a year but next time everyone will believe Sony ;)
 
It's like this blakjedi, as far as real flops go, the 360 isn't going to be pushing the terraflop barrier. If Allard and the rest of the fun-bunch start pushing it as a 1TFlpo + plus system, they're within their rights to do so since they can whip out their 'fuzzy math' calculator.

Now, if when Sony launches they go with the same 1TFlop+ crap, it'll be because they went down the same fuzzy math route as Microsoft to put themselves on an even footing. But the PS3 won't be having a real 1TFlops at it's disposal either, because that was all about the 'Broadband Engine,' and they're not getting that - they're getting a straight 1/4 of that if we're lucky.
 
You could draw a parallel to what Ken said regarding process size, sure he'd love 65nm, but you can't make a machine on dreams. 90 is what they're stuck with and i'm sure it'll be a a decent piece of hardware. Still there's gotta be some lil demon in people's heads wondering if they've missed something, the lack of hardware AA made the first generation PS2 games look pretty damn ugly.
 
xbdestroya said:
It's like this blakjedi, as far as real flops go, the 360 isn't going to be pushing the terraflop barrier. If Allard and the rest of the fun-bunch start pushing it as a 1TFlpo + plus system, they're within their rights to do so since they can whip out their 'fuzzy math' calculator.

Now, if when Sony launches they go with the same 1TFlop+ crap, it'll be because they went down the same fuzzy math route as Microsoft to put themselves on an even footing. But the PS3 won't be having a real 1TFlops at it's disposal either, because that was all about the 'Broadband Engine,' and they're not getting that - they're getting a straight 1/4 of that if we're lucky.

Agreed. Thats all I was saying... 8) [/i]
 
Thanks for the reply. So basically your saying that it would be impossible for MS, IBM & ATI to design any sort of system where by the messured output of the machine would reach 1Tflop with the tech as it is now without some kind of NVflops type thing. Even if they work together to design two chips that interact with each other very well & MS highly optimised system code?
Maybe I'm thinking very simply but i think it could be possible.

By the way, which is more processor intensive, geometry or shader computation. If its geometry I have an idea how they maybe able to get 1Tflop.
 
vblh said:
Thanks for the reply. So basically your saying that it would be impossible for MS, IBM & ATI to design any sort of system where by the messured output of the machine would reach 1Tflop with the tech as it is now without some kind of NVflops type thing. Even if they work together to design two chips that interact with each other very well & MS highly optimised system code?
Maybe I'm thinking very simply but i think it could be possible.

By the way, which is more processor intensive, geometry or shader computation. If its geometry I have an idea how they maybe able to get 1Tflop.

Depends how much they were willing to loose on the box.
They could probably build something that did massively wide vector operations that would also achieve it at the expense of dreadful performance in any useful app.

The point is it's not practical.
 
The only way to get 1 teraflop of programmable performance is a new design of chip that's quick on maths and scalable. That's why Cell was invented. If it could be attained other ways, it would have been. And even then you need several Cell chips to reach it. The cost is prohibitive.

In summary, again, it's not going to happen!

Plus you mentioned a requirement for MS to provide highly optimised system code, and MS aren't reknowned for highly optimised code! :p
 
function said:
xbdestroya said:
It's like this blakjedi, as far as real flops go, the 360 isn't going to be pushing the terraflop barrier.

What's a fake flop?

Flop numbers that include the graphics core in my mind, aren't fake, they're just not real. ;)

Saying that XBox 360 has over 1 TFLop of 'combined system performance' literally means nothing to me. That's like XBox's hundred+ whatever of GFlops performance, when the CPU was pushing upwards of 3.

Maybe I should have qualified it with, those are NVflops, rather than implying they weren't real. But really, is there a difference?
 
Titanio said:
For the last time - GPUs lend themselves very well to ultra high (inflated?) flop ratings.
So is Cell -- both are massively parallel processors.

It's a two-way street, both numbers are meaningless.

Sony and MS will both massively inflate FLOPS. I'd put more money on the MS solution being more balanced and useful in everything but marketing.
 
Status
Not open for further replies.
Back
Top