Console CPU's ?

Mythos

Regular
What are your thoughts on the final layout for each consoles cpus?

Sony Cell-

Will 65NM or 90NM make difference on the final PS3 cell. If so, what can it mean higher core speed-

" The first version of the chip will run at speeds faster than 4GHz. Engineers were vague about how much faster, but reports from design partners say 4.6GHz is likely. "

If 65nm, is it possible to go with 2 cell chips?

Or could it be some funky cell like chip 1pe+4APUs + 1PE+4apus, 1PE/6apu's?

Xenon-

Will their final chip be on 65NM or 90NM?

Will this be an effect on what will finally go in the console?

Will it be like leaked specs 3 cpu's or possibly 2?

Will it be different altogether; lower clock? Something like a modified G5 or G6 PowerPC?

Possibly something similar to what is done with the Cell PPE?

Ninetendo-

Since info is vague, make some predictions-

Will the CPU be on 65NM or 90NM?

Will this be an effect on what will finally go in the console?

Will it be more powerful than Xenon?

Will it be 1 CPU or multiple CPU's?

Will the speed clock be high or low?
 
I will say the ps3 cpu will be at 90nm 1pe x 8apus at 4ghz

xenon/xbox2 will have 2 cpus , 3 cores on each . I'm thinking a pretty big revamp of the power 5s


the new nintendo system will have 8 g5s running at 3 ghz each 1 gig of ram and 3 r600s :)
 
jvd said:
I will say the ps3 cpu will be at 90nm 1pe x 8apus at 4ghz

xenon/xbox2 will have 2 cpus , 3 cores on each . I'm thinking a pretty big revamp of the power 5s blah blah

Nice, one of the mods trolls his own forum... :p
 
Troll how ?

This is the newest info i've heard . 2 tri core cpus for the xenon and a 1x8 4ghz first quater 06 ps3 on 90nms

IF you think that is trolling you got another thing coming to you
 
It will blow my friggin mind if the CPUs are anywhere near the specs you just listed. Exactly how much money do you think these manufacturers want to lose per console?

What about cooling, eh? Are we going to have 50lb copper-cooled behemoths with heat pipes looking like exhaust pipes this gen?

There are a lot of people who are going to be let down by their crazy dreams me thinks......

Heh. I'd say 1 tri-core CPU for Xenon and a buggy, first-gen Cell CPU for PS3, with both equaling or slightly besting the top end desktop CPUs available today. And even that may be more than will happen because they aren't likely to drop a CPU that puts out over 100W of heat into a console.
 
JvD, I thought it obvious you were being intentionally silly in your post by suggesting SIX G5s for nextbox and just one cell for PS3. Particulary your last statement about 8 CPUs for revolution marked your post as obvious bullshitting.

That you now claim to have been serious makes it even more disappointing, because we've been over this a million times already. Nextbox isn't going to have G5 CPUs, we KNOW this! Particulary not 2x3 of them. Not only are they huge, they burn like 90W of power each at 2.5GHz.

So if your post was not meant to troll, you definitely put your foot in your mouth writing it. I don't know what kind of people you claim to "hear" your stuff from, but I wouldn't listen to them if I were you, they never seem to get their facts right, know what I'm saying here?
 
actually the power 5 was a type o , but the nintendo line was obvously a joke .

And actually the latest rumors on the ps3 is something else but if i posted it i would really get called a troll :oops:
 
jvd said:
actually the power 5 was a type o , but the nintendo line was obvously a joke .

And actually the latest rumors on the ps3 is something else but if i posted it i would really get called a troll :oops:

I dunno man, your 'rumors' haven't been very accurate before. ;)
 
yea sometimes they are bad , but they are just what i'm hearing .

Anyway as a serious ns5 spec

single g5 modified with more L3 cache , a custom graphics chip from ati heavly modified from ati's next next gen graphics chip (next gen being r520 :- )
 
My prediction:

Xbox 2 = 1 Tricore cpu @3.5 GHz
PS3 = Modified Cell 1pe +4 SPU's @ 4GHz
Nintendo = 1 dual core power PC CPU @ 3.5 GHz
 
Qroach said:
My prediction:

Xbox 2 = 1 Tricore cpu @3.5 GHz
PS3 = Modified Cell 1pe +4 SPU's @ 4GHz
Nintendo = 1 dual core power PC CPU @ 3.5 GHz

Why a 'modified' Cell? This is what Cell was meant to do! Also I think that although at some point Sony was probably thinking about a 1 PE + 4 SPE Cell chip for PS3, they'll try to get that extra 4 SPE's in there to make sure they match/exceed MS - either that or maybe go a 2 PE + 8 SPE route (ie 2x1PE/4SPE chips).
 
jvd said:
xenon/xbox2 will have 2 cpus , 3 cores on each . I'm thinking a pretty big revamp of the power 5s

I have actually heard the same thing... and i dont know jvd or who he knows...
 
xbdestroya said:
Also I think that although at some point Sony was probably thinking about a 1 PE + 4 SPE Cell chip for PS3, they'll try to get that extra 4 SPE's in there to make sure they match/exceed MS - either that or maybe go a 2 PE + 8 SPE route (ie 2x1PE/4SPE chips).
What's the point of having 2 PPEs? (Besides, check out the old thread)
 
one said:
xbdestroya said:
Also I think that although at some point Sony was probably thinking about a 1 PE + 4 SPE Cell chip for PS3, they'll try to get that extra 4 SPE's in there to make sure they match/exceed MS - either that or maybe go a 2 PE + 8 SPE route (ie 2x1PE/4SPE chips).
What's the point of having 2 PPEs? (Besides, check out the old thread)

Well don't worry, I'm no stranger to the Cell's die size and cost-related issues. The reason I think they might go for it though would be a) stronger performance in more generalized apps (ie apps in which the SPE's could not be brought to bear), and b) if indeed Sony is having issues getting 8 SPE chips, as has been rumored, it might make sense for them to just go ahead and bite the bullet to get those extra SPE's in there, plus the benefit of that extra PE.

I'm not saying any of the above is a likely scenario or anything, I'm just aying that since they've gone ahead and invested billions in Cell, granted for other applications above an beyond the PS3 as well, it would seem ridiculous to me for them NOT to ensure a configuration inside the PS3 that beats XBox 2's. And that means more than just flops IMO.
 
xbdestroya said:
a) stronger performance in more generalized apps (ie apps in which the SPE's could not be brought to bear),
What type of application requires more floating point performance than games?
xbdestroya said:
b) if indeed Sony is having issues getting 8 SPE chips, as has been rumored
I don't know that rumor and why not 6 instead of 4 when the Cell's die is designed as a rectangle to add/remove 2 SPEs at a time, so I asked.
 
one said:
xbdestroya said:
a) stronger performance in more generalized apps (ie apps in which the SPE's could not be brought to bear),
What type of application requires more floating point performance than games?

Well I'm not talking in terms of flops per se, because in that arena I'm not looking at the PE in and of itself as a significant factor. However, the PE's are able to do some things that the SPE's, as specialized as they are, are really not equipped to handle. In this sense I'm looking at the abilities of these consoles beyond gaming, and perhaps in the games themselves. I should add, I'm not a dev or anything, so maybe I underestimate the number of tasks the SPE's can be utilized for. I just feel that though diminished in flops performance, there is a consolation-prize sort of advantage held by the multiple cores fo the XBox 2 by virtue of their greater (slightly) versatility.

one said:
xbdestroya said:
b) if indeed Sony is having issues getting 8 SPE chips, as has been rumored
I don't know that rumor and why not 6 instead of 4 when the Cell's die is designed as a rectangle to add/remove 2 SPEs at a time, so I asked.

Well I'm not sure what we're talking about here, so it could be that I never did. Are you saying it doesn't have to be either 4 or 8, they could make it 6 if 8 was too difficult? I guess so - I don't see why not. It would be a good deal more cost effective than my extra PE theory, that's for sure.
 
xbdestroya said:
Well I'm not talking in terms of flops per se, because in that arena I'm not looking at the PE in and of itself as a significant factor. However, the PE's are able to do some things that the SPE's, as specialized as they are, are really not equipped to handle. In this sense I'm looking at the abilities of these consoles beyond gaming, and perhaps in the games themselves. I should add, I'm not a dev or anything, so maybe I underestimate the number of tasks the SPE's can be utilized for. I just feel that though diminished in flops performance, there is a consolation-prize sort of advantage held by the multiple cores fo the XBox 2 by virtue of their greater (slightly) versatility.

I'm curious. What kind of computing CAN'T a SPE do? I mean, I know that it really shines in math intensive applications, but otherwise, all processors do is basically do very simple math and move information from one place to another. So I'm curious what are the real limitations of SPE's in normal day to day applications, being that they are 99% of the times extremelly non-cpu intensive, having only I/O as their bottlenecks.
 
Alejux said:
xbdestroya said:
Well I'm not talking in terms of flops per se, because in that arena I'm not looking at the PE in and of itself as a significant factor. However, the PE's are able to do some things that the SPE's, as specialized as they are, are really not equipped to handle. In this sense I'm looking at the abilities of these consoles beyond gaming, and perhaps in the games themselves. I should add, I'm not a dev or anything, so maybe I underestimate the number of tasks the SPE's can be utilized for. I just feel that though diminished in flops performance, there is a consolation-prize sort of advantage held by the multiple cores fo the XBox 2 by virtue of their greater (slightly) versatility.

I'm curious. What kind of computing CAN'T a SPE do? I mean, I know that it really shines in math intensive applications, but otherwise, all processors do is basically do very simple math and move information from one place to another. So I'm curious what are the real limitations of SPE's in normal day to day applications, being that they are 99% of the times extremelly non-cpu intensive, having only I/O as their bottlenecks.

The limitations are simple enough that it can only runcode and read data directly from it's own local memory, so if you exceed 256Kb then you have to break the task down and if you can't do that then it won't run.

Trivial example might be an interpretted/GC'd language, most of which require a much larger footprint than 256K to run effectively.

If your trying to hide DMA latency the actual usable amount of memory would be less than that (typically 1/2 for most simple solutions).
 
ERP said:
The limitations are simple enough that it can only runcode and read data directly from it's own local memory, so if you exceed 256Kb then you have to break the task down and if you can't do that then it won't run.
maybe you would rephrase that as "it can only efficiently run code and data directly from it's own local memory".
In fact it seems, from preliminal informations, there's nothing a SPE can't do compared to a general purpose processor.

If your trying to hide DMA latency the actual usable amount of memory would be less than that (typically 1/2 for most simple solutions).
When you're doing calculations in streaming fashion you don't need huge buffers for your streamed data. Most of the local memory can be used to store constant data/code if it's needed.
If I were coding a simple T&L engine on a SPE probably I wouldn't need much more than a couple of 16KB buffers for incoming and outcoming data.
One thing is sure, SPEs are not your (and mine) AI dream-processors, but SPEs are not designed to be that and workarounds can be found anyway.
IMHO a SPE can efficiently run 80-90% of common game code, we have a PPE for everything else.
 
ERP said:
The limitations are simple enough that it can only runcode and read data directly from it's own local memory, so if you exceed 256Kb then you have to break the task down and if you can't do that then it won't run.
Well, IIRC PPE puts a task into an SPE, then the SPE on its own can DMA-transfer data from/to the main memory to/from Local Store, doing the task till its end. To scan a large data in the main memory, it seems possible that you have a small buffer in Local Store and stream data from the main memory to the buffer to process it with an SPE. Is it unusable in an actual scenario?
 
Back
Top