Well, Well, Xenon CPU and Cell the same??

[russo121]

I've found this article ( http://arstechnica.com/news.ars/post/20050308-4685.html) from Hannibal very interesting....
"...Xenon's three PPC cores are the same as the Cell's single PPC core..."

This is his opinion and worths a read!

 

[Titanio]

Been suspected by some for a while now, and wouldn't be very surprising at all.

 

[aaaaa00]

Last I heard there was some extra "special sauce" on the xenon PPC cores compared with the CELL PPC.

 

[j^aws]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

 

[russo121]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

Oh yeah?!?! So why everybody thinks PS3 will be faster and better than Xenon?

 

[hey69]

because Ken Kutaragi says so! it has the cell, it has the emotion

 

[Carl B]

I'm not saying Hannibal is wrong - I have nothing but respect for that man - but HOW can they be the same? Would Sony have allowed that? Wasn't the PPC core each is using a custom job? Something doesn't make sense. Either that or one of these companies ended up paying for the development of the other company's PPC.

 

[Carl B]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

Oh yeah?!?! So why everybody thinks PS3 will be faster and better than Xenon?

Because of the SPE's of course.

 

[Titanio]

I'm not saying Hannibal is wrong - I have nothing but respect for that man - but HOW can they be the same? Would Sony have allowed that? Wasn't the PPC core each is using a custom job? Something doesn't make sense. Either that or one of these companies ended up paying for the development of the other company's PPC.

The Power core is a relatively small player in the whole Cell picture. To give you an idea of what I mean, a lot of people don't even include it in performance calculations

The SPEs are Cell's workhorse.

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

Oh yeah?!?! So why everybody thinks PS3 will be faster and better than Xenon?

See above (although nothing's absolutely positively certain at this stage ).

 

[j^aws]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

Oh yeah?!?! So why everybody thinks PS3 will be faster and better than Xenon?

There are plenty of threads if you do a search on these forums but I'll just do a quick calc why,

The following units are components of that make up the programmable, 32bit single precision FLops,

Assuming FMADD ops,

VMX ~ 8 Flops per cycle
FPU ~ 2 Flops per cycle
PPE ~ VMX + FPU~ 10 Flops per cycle
SPE ~ 8 Flops per cycle

Xenon CPU core = VMX + FPU ~ 10 Flops per cycle

3 Cores ~ 30 Flops per cycle

3 Cores @ 3 GHz ~ 30* 3Ghz~ 90 GFlops



CELL CPU core = PPE + 8*SPE ~ 10 + 8*8 ~ 72 Flops per cycle

CELL @ 4Ghz ~ 72*4Ghz~ 296 GFlops

Now I expect the CELL in PS3 to be less powerfull than the above. This excludes the power of their respective GPUs but the expectation of why people are thinking this is because of the CPUs difference in Flops. This has been speculated for months here and is nothing new. The final output on screen may not be that different in the end. Who knows until official info is released...

 

[Carl B]

I'm not saying Hannibal is wrong - I have nothing but respect for that man - but HOW can they be the same? Would Sony have allowed that? Wasn't the PPC core each is using a custom job? Something doesn't make sense. Either that or one of these companies ended up paying for the development of the other company's PPC.

The Power core is a relatively small player in the whole Cell picture. To give you an idea of what I mean, a lot of people don't even include it in performance calculations

The SPEs are Cell's workhorse.

Still though, it doesn't jive with me. Obviously those PPC cores are good enough to be the basis of MS's entire console, nevermind the SPE's. Is this a generic 'gaming' core IBM came up with and sold to Sony and MS? Is Nintendo going to be using the same core as well? Did the PPC teams working on both the Cell and the Microsoft chips come to the same conclusions seperately?

I don't care that it's a small part of Cell - it just seems like insanity from a business perspective. Where were the corporate firewalls?

If I were Sony and I found out that the custom PPC/PE core I had helped develop and paid for was now basically the entire basis of Micrtosoft's new console, I would want answers.

It is SO egregious, it makes me wonder if IBM just didn't come up with some generic PPC design for gaming and sell it to all three console makers. That seems to me the only situation that would make sense to me. And if that were that case, I don;t know, it just seems kind of lazy on all the manufacturers parts - though especially Sony, who was already their investing the billions.

 

[MechanizedDeath]

I'm not saying Hannibal is wrong - I have nothing but respect for that man - but HOW can they be the same? Would Sony have allowed that? Wasn't the PPC core each is using a custom job? Something doesn't make sense. Either that or one of these companies ended up paying for the development of the other company's PPC.

The Power core is a relatively small player in the whole Cell picture. To give you an idea of what I mean, a lot of people don't even include it in performance calculations

The SPEs are Cell's workhorse.

Still though, it doesn't jive with me. Obviously those PPC cores are good enough to be the basis of MS's entire console, nevermind the SPE's. Is this a generic 'gaming' core IBM came up with and sold to Sony and MS? Is Nintendo going to be using the same core as well? Did the PPC teams working on both the Cell and the Microsoft chips come to the same conclusions seperately?

I don't care that it's a small part of Cell - it just seems like insanity from a business perspective. Where were the corporate firewalls?

If I were Sony and I found out that the custom PPC/PE core I had helped develop and paid for was now basically the entire basis of Micrtosoft's new console, I would want answers.

It is SO egregious, it makes me wonder if IBM just didn't come up with some generic PPC design for gaming and sell it to all three console makers. That seems to me the only situation that would make sense to me. And if that were that case, I don;t know, it just seems kind of lazy on all the manufacturers parts - though especially Sony, who was already their investing the billions.

If they didn't use a whole processor, they could just cut and paste parts of a processor. I don't see a problem with this, it's not like Cell and the XeCPU are identikit chips. For the XeCPU, the PPC core is the heart of the chip, and doing most of the work. For the Cell, the heart is the SPEs, so maybe the PPC was just a convenient processor to use for its specific role. IBM's allowed to borrow ideas from its internal teams. PEACE.

 

[aaaaa00]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

That's not the only "special sauce" I heard about.

 

[Carl B]

If they didn't use a whole processor, they could just cut and paste parts of a processor. I don't see a problem with this, it's not like Cell and the XeCPU are identikit chips. For the XeCPU, the PPC core is the heart of the chip, and doing most of the work. For the Cell, the heart is the SPEs, so maybe the PPC was just a convenient processor to use for its specific role. IBM's allowed to borrow ideas from its internal teams. PEACE.

Didn't know they were able to borrow from internal teams - I thought that the Cell project was closed off. Oh well whatever, I'll ask some friends of mine at IBM how they think it came about. Still, just seems strange.

I don't disagree that the heart of Cell is the SPE's, I'm just wondering who's dollars it was that paid for the PPC core development. Maybe that was straight-up IBM.

 

[Carl B]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

That's not the only "special sauce" I heard about.

Well you can't just leave it at that.

 

[archie4oz]

Wouldn't be a surprise at all as it's been speculated here for several months. Though the Xe cores will likely have fatter registers, 128.128 bit, and custom instructions unique to Xenon compared to the CELLs PPE.

I'm not sure that could really even be considered "special sauce"... If it's still VMX then it's the same as always, with 32 logical registers, and the 128 being physical (which isn't a big deal since my 970 itself has like 80 physical VMX registers)...

 

[Megadrive1988]

Sega's Model 3 board used 1 PowerPC 603e

while 3DO's M2 used 2 seperate PowerPC 602s

not the same processor.

I doubt very much that Xenon CPU cores are identical to the PU aka PPE, the PowerPC 'housekeeping' core in CELL. they probably share the same or similar technology base, or elements, but not exactly the same.

ArsTechnia article is just speculating. they don't know any more than anyone else that is not 'in-the-loop'

 

[aaaaa00]

I'm just wondering who's dollars it was that paid for the PPC core development. Maybe that was straight-up IBM.

IBM gets paid twice for the same R&D, what's not to love from IBM's perspective?

 

[Carl B]

I'm just wondering who's dollars it was that paid for the PPC core development. Maybe that was straight-up IBM.

IBM gets paid twice for the same R&D, what's not to love from IBM's perspective?

Well, if it was THEIR venture dollars performing that R&D, nothing's not to like. 8)

If they were commisioned by one of the other two to come up with that core, one of them has a lot not to like.

 

[s1lverbak]

VMX ~ 8 Flops per cycle
FPU ~ 2 Flops per cycle
PPE ~ VMX + FPU~ 10 Flops per cycle
SPE ~ 8 Flops per cycle

Xenon CPU core = VMX + FPU ~ 10 Flops per cycle

3 Cores ~ 30 Flops per cycle

3 Cores @ 3 GHz ~ 30* 3Ghz~ 90 GFlops



CELL CPU core = PPE + 8*SPE ~ 10 + 8*8 ~ 72 Flops per cycle

CELL @ 4Ghz ~ 72*4Ghz~ 296 GFlops

I guess I am confused...Everyone seems to be using the term "Core" pretty liberally. From the specs posted at gamespy, it says that each CPU would be able to do 2 instructions per cycle. Wouldnt that make it dual core? So it wouldnt be 3 Cores @ 3GHz, it would be 6 @ ?GHz. Right??