IBM Opens Cell Processor

[MacRumors]

http://www.macrumors.com/images/macrumorsthreadlogo.gif (http://www.macrumors.com)

eeTimes is reporting (http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=163106213) that IBM and its partners are planning on opening the upcoming Cell processor's architecture and provide software development libraries in an effort to increase adoption of the new architecture.

The Cell processor will be used in Sony's upcoming Playstation 3. It was announced in February (http://www.macrumors.com/pages/2005/02/20050207161449.shtml) by IBM, Sony and Toshiba and has always expected to be used in a broad range of applications - including video processing, medical imaging and high-performance computing.

There is no direct link to Apple, but with IBM's involvement, speculation abounds.

[Bubbasteve]

Man I hope they put this bad boy in the Macs soon

[qubex]

G6

[tazznb]

I really hope Apple jumps on this technology.

Hopefully IBM won't give processor speeds promised to Apple, renig, and then give it to game console makers.

BTW: Until we hear it directly from the horse's mouth NOBODY HERE KNOWS if the chips in the Xbox 360 were intended for Apple's computers.

If there is undeniable proof, and not hearsay please post it. :(

[JackSYi]

:eek: :eek: G6, I second that. :eek: :eek:

[Peace]

I've said it before! I think

Think G6 WWDC

[poundsmack]

i think if we saw CELL in macs it would only be in the servers. but who knows.

[thedoc1111]

Anyone who thinks the cell will be a G6 is completely wrong - it is like saying that because my NVidia 6800 is extremely powerful, I should replace my CPU with one of them.

The cell is extremely good at doing specific types of calculation, but appalling for running the code that we rely on in complicated Operating Systems and Applications.

However, the argument above does make sense if you swap out replace for augment - I wish Apple would ship computers with an accessory Cell processor to handle CPU intensive repetitive tasks such as H.264 decoding/encoding. We'll have to wait and see, though.

[amholl]

This CANNOT BE A G6!!!! OMG weve said it a million times!

Anyway, the best use for this chip, as it is a MULTIMEDIA CHIP AND IS NOT SUITABLE FOR NORMAL COMPUTING, is to be a co-proc and do something like offload H.264 tasks or video and rendering, but not have it be THE proc

[Cooknn]

Think G6 WWDCEh, the G6 is already here (http://www.pontiac.com/g6/index.jsp?brand=home) :p They'll have to come up with something else me thinks :o

[poundsmack]

what is there was a central processor, say like a G5, and them many CELL processors on the board to take over the tasks.

[iJaz]

Hmm, Cell chips, Intel rumors, new announced Xbox tripple core chips...
This can only mean Powerbook G5 on Tuesday! :p

[blitzkrieg79]

Hmmmmmmmm, isn't it funny how IBM "opened up" the CELL technology to others after the weekends Apple and Intel speculations?

Anyway I do hope that CELL-like (I say CELL-like because the PS3 CELL is not suited for regular computing BUT it does have potential to be one)...
Anyway, CELL technology seems to be versatile and adaptable to specific users needs, at least thats what I got from reading all the articles I found so far...

Its floating performance seems to be very good, integer performance can be improved by changing the PPE unit into something such as a current G5, I know the size of such a chip would be big but it wouldnt be so much bigger than current two 970fxs, and with performance gains of the CELL (especially in the floating point arena, and isn't that what Apple all those multimedia apps require?) I could care less about MS word and such, it already ran good enough on 1GHZ machines so integer performance doesnt matter a whole lot to me...
Apple really wouldnt have to use Dual configurations in PowerMacs, maybe create an Xstation for the multiprocessor configurations and thats it...

EDIT: And Cell is a specialized processor but not as specialized as lets say a GPU, CELL is capable of running regular routines too, anyway, like I said, just change the PPE of the CELL to a Power4-Power5 derivative, and as far as programming goes, you can't really expect programming to stay still when processor technology evolves, programming has to evolve too, anyway, Apple likes giant leaps and this would definately be one giant leap...

[Sharewaredemon]

This CANNOT BE A G6!!!! OMG weve said it a million times!

Anyway, the best use for this chip, as it is a MULTIMEDIA CHIP AND IS NOT SUITABLE FOR NORMAL COMPUTING, is to be a co-proc and do something like offload H.264 tasks or video and rendering, but not have it be THE proc

I just hope a computer comes out like what you have described for when I'm about to graduate (3 years) because that would be extremely useful for video editing.

[pgwalsh]

I wish Apple would ship computers with an accessory Cell processor to handle CPU intensive repetitive tasks such as H.264 decoding/encoding. We'll have to wait and see, though.
That would be interesting. I wonder how feasable that would be. Dual Processor G5 PowerMac with an additional cell processor for use with FCP and Logic.. Perhaps an add in PCI card with two cells for offloading. Would that even be possible? I suppose you'd have to rewrite your applications to utilize this applicaiton, but that would pretty slick.

[enygma]

Anyone who thinks the cell will be a G6 is completely wrong - it is like saying that because my NVidia 6800 is extremely powerful, I should replace my CPU with one of them.
Or not... considering sony saw it fit to include a GPU as well in their PS3 outside of the Cell processor. Cell is capable of general purpose computing, and IBM even goes and brags that they have used multiple operating systems on Cell. Heck, both sony and IBM do. At the Game Developers Conference, it was something they kind of highlighted on with 2 of their lectures that I had attended. The Cell processor is a capable processor. It is also a very powerful processor, and nowhere near what I would consider comparing to a GPU. I would compare it more to a dual core Opteron running at 3.2GHz... only this Opteron seems to have 8 additional AthlonXP processors around it(The SPEs) running at the same clock speed, all with their own cache and DMA channels, while at the same time, be able to punch the ammount of memory through per second that only 4 to 5 Opterons operating in NUMA would be able to do... in one chip.

[thedoc1111]

Perhaps an add in PCI card with two cells for offloading. Would that even be possible?

Almost there, but no cigar.

PCI does not have the bandwidth required, but PCI Express would, especially as PCI Express can be easily scaled, from a 1x port (for peripheral cards) to a 16x port (currently for graphics cards). A Cell could easily run on one of these cards, given that they supply up to 75W of power.

All you would need is tie-ins from the OS, which Apple could probably code in for 10.5 or a major update for 10.4

The only worry would be that this extra code could lead to security vulnerabilities...

[pgwalsh]

Or not... considering sony saw it fit to include a GPU as well in their PS3 outside of the Cell processor. Cell is capable of general purpose computing, and IBM even goes and brags that they have used multiple operating systems on Cell. I've read that, but I also read that they can be programmed through software to perform specific tasks. I would guess you could use it for compression or processing filters etc??

[thedoc1111]

Or not... considering sony saw it fit to include a GPU as well in their PS3 outside of the Cell processor. Cell is capable of general purpose computing, and IBM even goes and brags that they have used multiple operating systems on Cell. Heck, both sony and IBM do. At the Game Developers Conference, it was something they kind of highlighted on with 2 of their lectures that I had attended. The Cell processor is a capable processor. It is also a very powerful processor, and nowhere near what I would consider comparing to a GPU. I would compare it more to a dual core Opteron running at 3.2GHz... only this Opteron seems to have 8 additional AthlonXP processors around it(The SPEs) running at the same clock speed, all with their own cache and DMA channels, while at the same time, be able to punch the ammount of memory through per second that only 4 to 5 Opterons operating in NUMA would be able to do... in one chip.

Cell is capable of general purpose computing - but MUCH slower than our current designs. It's advantage is that, with specially optimised code it can run much faster, and cooler and cheaper, but only for vector code.

Games consoles (with the single exception of the original XBox) have always been run on customised, special purpose chips, and the PS range have always run on chips that specialise in vector processing. This is fine in game design, as game engines have traditionally been quite small and easily optimised, but not up to running bulky stuff like Word v.X

Edit: One final issue - all that DMA everywhere does make me extremely concerned for the security of these chips. I have heard that much of the processor's controllers is written in software reather than burned into the ASIC (which is what makes this specs release extremely important, btw) but still if a vulnerability exists within the silicon it could give the cell a 'not-production-ready' image for the high end market

[James Philp]

May I just remind everyone at this point of a thread I started not so long ago and got shot down for!:
http://forums.macrumors.com/showthread.php?t=120197
Bar-stewards!
Well all I can say now is :p :p :p :p

[enygma]

I've read that, but I also read that they can be programmed through software to perform specific tasks. That means you could use it as a dedicated processor for specific tasks.
This wouldn't surprise me as I believe they adopted some of the ideas used in their Blue Gene high performance clusters, which are reprogrammable hardware.

[modernpixel]

I don't think the Cell is quite so limited. I've read through this fantastic explanation:

http://www.blachford.info/computer/Cells/Cell1.html

It's the future of CPUs, Apple would be foolish not to work with IBM to modify it to make the fastest freakin' Mac ever (or PC for that matter.)

Screw the talks with Intel, they've already got an "in" to the next best thing.

Joe

[enygma]

Cell is capable of general purpose computing - but MUCH slower than our current designs. It's advantage is that, with specially optimised code it can run much faster, and cooler and cheaper, but only for vector code.
Link? Last bit of vector code I recall in use was for the VU0 and VU1 processors in the PS2. The standard processor in the PS2 used the MIPS architecture. Unfortunately, VU0 and VU1 had to be utilized using assembly, while Cell has been designed to be a much easier programming target using C++.

[pawnstar]

I don't think the Cell is quite so limited. I've read through this fantastic explanation:

http://www.blachford.info/computer/Cells/Cell1.html

It's the future of CPUs, Apple would be foolish not to work with IBM to modify it to make the fastest freakin' Mac ever (or PC for that matter.)

Screw the talks with Intel, they've already got an "in" to the next best thing.

Joe

interesting link dude, cheers :)

[poundsmack]

i dont think anyone of us can predict what is going to happen with this. but i can say this. only good will come out of it.

[iomar]

Yes, good! But we are moving to intel processers.

[Peace]

Sorry to imply the cell as the new G6..
I was just trying to say new G6 at the WWDC..
What else is left to announce there??
Tiger's out already.
FCP is out already etc..
The G5 was bumped to 2.7 a while back

Powerbook G5 ?
PowerMac G6 ?

It's part of the evolution process!!! :)

[ifjake]

sweet i get to join that club of people that posted something before it reaches the front page.

http://forums.macrumors.com/showthread.php?t=128562

:D

this is nice, but i would rather Apple develop something new for the notebook arena. the CELL seems like it'll pretty much be inhabiting a tower for some time. but i'm being narrow minded i know. this is still an exciting possibility.

[MacVault]

Why in the ---- would Apple not make use of this processor, assuming it's all that it's cracked up to be?

And why in the ---- is Apple putting along at 2.something Ghz when the new XBOX has an IBM PowerPC-based CPU with 3 symmetrical cores running at 3.2GHz each.

It's pissing me off! I remember when the G5 came out. Wooo-Hooo!!! It was suppose to be the holy-frickin-grail with great new architecture, etc. up to 3ghz after a few months etc. Well, instead of that we get a duldrum of 100 or 200 mhz updates at a time.

:mad: :mad: :mad:

WTF?!

[shamino]

Yes, good! But we are moving to intel processers.
Says who? Some random rumor page?

[shamino]

And why in the ---- is Apple putting along at 2.something Ghz when the new XBOX has an IBM PowerPC-based CPU with 3 symmetrical cores running at 3.2GHz each.
What makes you think the CPUs in an XBox are even close to the PPC 970 in terms of features or performance-per-clock?

Comparing a CPU that was custom-designed for XBox games to a general purpose workstation CPU is just silly.
It's pissing me off! I remember when the G5 came out. Wooo-Hooo!!! It was suppose to be the holy-frickin-grail with great new architecture, etc. up to 3ghz after a few months etc. Well, instead of that we get a duldrum of 100 or 200 mhz updates at a time.

:mad: :mad: :mad:

WTF?!
You're right. IBM has these amazing chips that outperform everything in the universe, but Apple thinks you don't want them, so they are refusing to use them in a computer. Which is why IBM is now selling them exclusively to game console makers.

And if you believe that, there's a room waiting for you at the Massachusettes State Home for the Bewildered.

[ifjake]

Yes, good! But we are moving to intel processers.

instigator!!

[Frisco]

Says who? Some random rumor page?

I think he was being sarcastic :rolleyes:

At least that's the way I took his comment--hope so.

[shawnce]

Why in the ---- would Apple not make use of this processor, assuming it's all that it's cracked up to be?
It depends on just how flexible it is and how much work it would take by Apple and 3rd party developers to get things running on par with today's G5 (PPC970xx) based systems. In other words are the gains worth it given the work it would take to achieve those gains and knowing that the G5 family (or its next generation) won't be standing still during this time.

The cell processor that will be used in PS3 has a limited PPC core in comparison to a G5. It would do poorly in comparison to a G5 on existing code streams and even with ones more coded to its model (it simply has less compute power then a G5).

The real power of the Cell processor is in the 8 on die specialized processing units (APU / SPE). Nothing in existing Mac OS X or from 3rd party developers has any concept of such things directly. To leverage them would take learning, new tool chains, and specific coding by developers.

In the case of Mac OS X at least Apple has a few technologies that could leverage APUs if coded to do so and if used those technologies in their applications they would see gains. The obvious examples are Core Audio, Core Image, and QuickTime codecs. Many aspects of those frameworks could be coded to use APUs (will take time for Apple to do this) while shielding developers that use those frameworks from having to know how the magic works.

And why in the ---- is Apple putting along at 2.something Ghz when the new XBOX has an IBM PowerPC-based CPU with 3 symmetrical cores running at 3.2GHz each.
What the Xbox 360 is using is different from a G5 in many ways. It is a simpler CPU with a limited execution model, hence much more dependent on compiler and/or customized coding to make it work well. Also it has fewer execution units then a G5. The G5 on the other hand is rather good at general computing and to do that comes with more a complex execution model.

In other words don't fall into the MHz/GHz myth... you cannot compare the computer power of different cores on clock rate alone.

[minimax]

This CANNOT BE A G6!!!! OMG weve said it a million times!

Anyway, the best use for this chip, as it is a MULTIMEDIA CHIP AND IS NOT SUITABLE FOR NORMAL COMPUTING, is to be a co-proc and do something like offload H.264 tasks or video and rendering, but not have it be THE proc

perhaps you could explain what exactly is a multimedia chip and why the cell is not suitable for normal computing? it has been said a million tiimes: a powerpc core with EXTRA FPU & INT co-processors. Yes, the PS3 cell is a stripped down Power4 / 5 processor but it is a SPECIALISED chip for the playstation, just like the G5 is a specialised PPC for the Macintosh platform. Also, the fact that it needs specialised software is nothing new. MMX, SSE, Altivec and even dualcore ask for software implementation, and according to this (http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2377) article writing optimised code for dual core is a real b*tch too, perhaps even more so then the CELL since there are full threads with multicore processing that need to interact with eachother. What i understand form that article dual processing can give an initial boost but if software developers don't succeed in developing efficient tools the multi-core promise may just be a Fata Morgana. Cell-processing seems to be a different story.

The IBM press release indicates the Cell processor is "Multi-thread, multi-core" but since the APUs are almost certainly not multi-threaded it looks like the PU may be based on a POWER5 core - the very same core I expect to turn up in Apple machines in the form of the G6 [G6] in the not too distant future
- http://www.blachford.info/computer/Cells/Cell1.html


If it's not multithreading (which seems pretty logical with one main processor) we are probably talking extensive Instruction Level Parallelism. ILP needs far less optimised software and already has proved it's potential.

[joeboy_45101]

Hopefully Apple will look into this amazing technology. I also hope that this will soften the relationship between Apple and IBM. The G5 development has been good, but we definitely should have been to at least 3Ghz by now.

I hate it when somebody portrays the G4 or G5's power in terms of a X Ghz G4 or G5 is equal to a 2X Ghz or < X Ghz Pentium 4. We honestly can't compare processor power like that. If Apple wants to convince people of the power of it's processor then it's going to have to be up against a processor of equal clock frequency.

Who wants to bet that a 3.2Ghz G5 would smoke a 3.2Ghz Pentium 4?

[nuckinfutz]

Who wants to bet that a 3.2Ghz G5 would smoke a 3.2Ghz Pentium 4?

Hell a 2.7GHz G5 will smoke a 3.2Ghz P4

I'd like to see Apple embrace Cell where it offers a tangible improvement over other chips. Encoding h.264 is a bear. We needs some massive glops to toss at this "pain"

[griz]

I'm beginning to think that this site should be called MacSpeculation.com
Any use of the CELL in Apple machines is speculative at best. Can anyone show some reasonable source for Apple using Cell other than their own hopes and dreams of a next generation Powerbook or Powermac?

I would love to see Apple come out with something that just trumps everyone and I am guessing that they are working on it. I can't imagine Apple moving up 200Mhz at a time for the next few years. They need a bump like the 68k to PPC move. They successfully pulled off the 68k to PPC move and then the OS9 to OSX move. It's time for them to put the lid on X86 architecture once and for all and kill it with something over the top. As for what that is? It's anyone's guess.

[BornAgainMac]

What would happen if a Cell processor would be used as a general computer and integer math? How much slower compared to let's say a G3 or Pentium III? I want my video encoding to be faster. It could be a third processor in my G5 for those tasks.

Perhaps have a Cell Mini that it functions as the only processor for simple web browsing and video encoding.

[isgoed]

Or not... considering sony saw it fit to include a GPU as well in their PS3 outside of the Cell processor. Cell is capable of general purpose computing, and IBM even goes and brags that they have used multiple operating systems on Cell. Heck, both sony and IBM do. At the Game Developers Conference, it was something they kind of highlighted on with 2 of their lectures that I had attended. The Cell processor is a capable processor. It is also a very powerful processor, and nowhere near what I would consider comparing to a GPU. I would compare it more to a dual core Opteron running at 3.2GHz... only this Opteron seems to have 8 additional AthlonXP processors around it(The SPEs) running at the same clock speed, all with their own cache and DMA channels, while at the same time, be able to punch the ammount of memory through per second that only 4 to 5 Opterons operating in NUMA would be able to do... in one chip.
Cell is capable of general purpose computing - but MUCH slower than our current designs. It's advantage is that, with specially optimised code it can run much faster, and cooler and cheaper, but only for vector code.
Link? Last bit of vector code I recall in use was for the VU0 and VU1 processors in the PS2. The standard processor in the PS2 used the MIPS architecture. Unfortunately, VU0 and VU1 had to be utilized using assembly, while Cell has been designed to be a much easier programming target using C++.As a sidemark let me remark that the programming language says little about the architecture.
You wanted validation of thedoc1111's claim. Well anantech says: (see http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2379&p=3 )

There's not much that's impressive about the PPE, other than it's a small, fast, efficient core. Put up against a Pentium 4 or an Athlon 64, the PPE would lose undoubtedly, but the PPE's architecture is one answer to a shift in the performance paradigm. Performance in business/office applications requires a very powerful, very fast general purpose microprocessor, but performance in a game console, for example, does not. The original Xbox used a modified Intel Celeron processor running at 733MHz, while the fastest desktops had 2.0GHz Pentium 4s and 1.60GHz Athlon XPs. Given that the first implementation of Cell is supposed to be Sony's Playstation 3, the simplicity of the PPE is not surprising. Should Cell ever make its way into a PC, the PPE would definitely have to be beefed up, or at least paired with multiple other PPEs.and


Each SPE is only capable of issuing two instructions per clock, meaning that at best, each SPE can execute two instructions at the same time. The issue width of a microprocessor can determine a big part of how large the microprocessor will be; for example, the Itanium 2 is a 6-issue core, so being a 2-issue core makes each SPE significantly smaller than most general purpose microprocessors. Obviously somewhat more conservative than your claims. Furthermore there is a real good in depth loook at the cell on arstechnica and he comes with his own insight on apple's business strategy: (from: http://arstechnica.com/articles/paedia/cpu/cell-2.ars )
The Cell and Apple

Finally, before signing off, I should clarify my earlier remarks to the effect that I don't think that Apple will use this CPU. I originally based this assessment on the fact that I knew that the SPUs would not use VMX/Altivec. However, the PPC core does have a VMX unit. Nonetheless, I expect this VMX to be very simple, and roughly comparable to the Altivec unit o the first G4. Everything on this processor is stripped down to the bare minimum, so don't expect a ton of VMX performance out of it, and definitely not anything comparable to the G5. Furthermore, any Altivec code written for the new G4 or G5 would have to be completely reoptimized due to inorder nature of the PPC core's issue.

So the short answer is, Apple's use of this chip is within the realm of concievability, but it's extremely unlikely in the short- and medium-term. Apple is just too heavily invested in Altivec, and this processor is going to be a relative weakling in that department. Sure, it'll pack a major SIMD punch, but that will not be a double-precision Alitvec-type punch.Sure the cell can find great appliances in the mac, but it indeed depends very much on what exactly those appliances are to say whether the cell will shine at it.

[Lord Kythe]

Who wants to bet that a 3.2Ghz G5 would smoke a 3.2Ghz Pentium 4?

Gee... This is a tough one... I don't know... The P4 is SOOO powerful, and oh my SOOO quick... It's a shame it can't use more than one of those GHz while the rest of the CPU "potential" is busy fumbling with inefficient code...

I'd go with the G5 providing we ever get there before it is replaced by a POWER5 or 970MP derivative. P4s are already at 3.6 GHz without liquid-cooling. Obviously PCs being able to use that chip is only a "best-case-scenario-running-under-Mac-OS-X-PC-edition" kind of thing, but let's give Intel just a little bit of credit for upping the ante for IBM. Apple can preach about the MHz myth all they want, people are people and your average PC user knows only this: the higher the better.

Apple: you want to invest in advertising? GET HIGHER CPU CLOCK SPEEDS! Those PC users won't have ANYTHING to say for their Pentium 5 4.2 GHz against a G5/G6 5 GHz, and will switch for the bragging rights.

Just my 2 cents...

[Mitthrawnuruodo]

To those thinking the Cell is not suitable for "normal" computing: IBM and Sony had a working prototype Cell-powered workstation back in November 2004. (Link... (http://www.sony.net/SonyInfo/News/Press/200411/04-1129BE/)) I see no reason why Apple shouldn't be thinking about doing the same. :rolleyes:

A Cell-based (G6 or whatever they want to call it) Mac showcased at WWDC would be awesome, and another advantage is that the Cell seems to scale better than the Power4 (which the G5 is based on) so maybe we'll see 64 bit portable Macs not too far into the future. :D

[Dr.Gargoyle]

Hmm, Cell chips, Intel rumors, new announced Xbox tripple core chips...
This can only mean Powerbook G5 on Tuesday! :p
I didn't knew Apple was about to expand into the BBQ market....amazing ;)

[Dr.Gargoyle]

Sorry to imply the cell as the new G6..
I was just trying to say new G6 at the WWDC..
What else is left to announce there??
Tiger's out already.
FCP is out already etc..
The G5 was bumped to 2.7 a while back

Powerbook G5 ?
PowerMac G6 ?

It's part of the evolution process!!! :)
What about
iTMS selling movies?
iPod video/tablet?
dual core G4 PB?
two button mouse with scroll wheel?
the list goes on and on and on....

[miketcool]

I can see this now, 2 Dual Core PowerPC chips, a dedicated Cell video processor and an advanced GPU. We can start replacing our furnace with this advanced space heater!

[Dr.Gargoyle]

....And if you believe that, there's a room waiting for you at the Massachusettes State Home for the Bewildered.
LMAO... be nice will ya. I can feel his frustrartion. But admit it, it would be cool with a laptop game consol.
The brand new PBGC G6....doubling as a BBQ... :eek: :D

[blitzkrieg79]

As a sidemark let me remark that the programming language says little about the architecture.
You wanted validation of thedoc1111's claim. Well anantech says: (see http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2379&p=3 )
and

Obviously somewhat more conservative than your claims. Furthermore there is a real good in depth loook at the cell on arstechnica and he comes with his own insight on apple's business strategy: (from: http://arstechnica.com/articles/paedia/cpu/cell-2.ars )
Sure the cell can find great appliances in the mac, but it indeed depends very much on what exactly those appliances are to say whether the cell will shine at it.

So why can't the "primitive" PPE unit found in the current CELL processor of PS3 be replaced with a 970fx or power5 derivative?
And also why can't a CELL act as a coprocessor for multimedia intensive apps?
I really don't see any reasons why it can't do either...
Anyway, the SPUs of CELL are vector units (INDEPENDENT vector units, thats what makes them special), in the G4-G5 design Altivec is the vector unit but it can't act alone... So theoretically, Take a Power5 derivative (I know its nonexistant but who knows what IBM/Apple is working on right now) and attach a couple of Altivec2 (yes I know its non existant either but there have been some rumors about it) units and make them independent and how would that compare to a CELL...

EDIT: Come to think of it, I wonder how different will be the SPU units of the CELL compared to Altivec2

[Lord Kythe]

Jobs said a few days ago (if that) that Apple was not changing its course of actions. Why would he switch from PPC to a gaming-console customized chip (the Cell)?

When did power mac users envy the processors in gaming consoles? (might I add, which specs have yet to be confirmed)

I believe in PPC chips, they have always been good for Apple, and now that more and more companies are recognizing PPC (triple PPC core in XBox 360 and Nintendo's new Revolution for example) chips, you want me to believe the Cell might even be considered?

Developers are having enough of a hard time already optimizing their software for all that the G5 can offer (including AltiVec), I don't think Apple would add to the monumental task by replacing or even simply adding a Cell processor for specific tasks.

970MP, POWER5 derivative or even a totally new PPC chip, I can see.
The Cell or any direct derivative, I can't. Sorry. Don't see it.

You want a gaming console? Go to Walmart. You want a multi-purpose computer? Go to the Apple Store.

[minimax]

Jobs said a few days ago (if that) that Apple was not changing its course of actions. Why would he switch from PPC to a gaming-console customized chip (the Cell)?

When did power mac users envy the processors in gaming consoles? (might I add, which specs have yet to be confirmed)

I believe in PPC chips, they have always been good for Apple, and now that more and more companies are recognizing PPC (triple PPC core in XBox 360 and Nintendo's new Revolution for example) chips, you want me to believe the Cell might even be considered?

Developers are having enough of a hard time already optimizing their software for all that the G5 can offer (including AltiVec), I don't think Apple would add to the monumental task by replacing or even simply adding a Cell processor for specific tasks.

970MP, POWER5 derivative or even a totally new PPC chip, I can see.
The Cell or any direct derivative, I can't. Sorry. Don't see it.

You want a gaming console? Go to Walmart. You want a multi-purpose computer? Go to the Apple Store.

On the one hand you state the CELL is a gaming-console costumized chip, therefore it cannot be used in a Macintosh, on the other hand you point to the capabilities of the 'PPC' since it is being used in gaming-consoles. Not only is that a strange argument, but I would recommend to read some more on the CELL and PPC before making such statements, as the CELL is a customized PPC just like the XBOX360 processor but one with additional potential beyond the gaming console.

[fluidinclusion]

What about
iTMS selling movies?
iPod video/tablet?
dual core G4 PB?
two button mouse with scroll wheel?
the list goes on and on and on....


I'll give you the fact that there is something to the first three, but for an entire conference to even partially revolve around a new mouse is hilarious! NOT.

[pawnstar]

Apple have moved graphics over to core video/graphics - and core data etc for the the other elements. In effect they are abstracting the the system from the hardware - for Apple to shift over to something else eg Cell, it would appear they would only have to write code inside the system. Apps would be able to take advantage of the power afforded by the extra chips by default.

This is also similar to AltiVec (back in the day) - PS runs without it, but slower. Similar to 'ripples' working or not in Dashboard today with different GPUs.

The Cell chips are designed to work together. Higher end Macs could have lots and lower end could have fewer. This way apps could take advantage (a la AltiVec) and for the high end run faster with the greater workloads imposed on them.

[pawnstar]

maybe we'll be able to run PS3 games on our macs as suggested here:

http://www.mobilegamefaqs.com/newsstory.php?id=121

:)

[wdlove]

Let's just hope that Apple will adopt the Cell. Looking toward the G6 would be great. It will certainly be a post 970 chip.

[oskar]

Hmm, Cell chips, Intel rumors, new announced Xbox tripple core chips...
This can only mean Powerbook G5 on Tuesday! :p

Why do people keep doing this? :(

[Dr.Gargoyle]

I'll give you the fact that there is something to the first three, but for an entire conference to even partially revolve around a new mouse is hilarious! NOT.
Since when have Jobs just announced one thing at any of his keynotes? :rolleyes: He will need something for the "just one more thing....".
My guess is that Apple will just present more than one thing, since none of the above could be considered as jaw dropping.

[oskar]

Duplicate post, sorry.

[oskar]

People say that this isn't happening because the cell has specific purposes like multimedia. Last time I checked iMovie, iPhoto, iDVD and iTunes were multimedia apps! How about FCP, Motion, DVDSP, Soundtrack, Logic, Shake, etc. Apple's main apps. Even Tiger has multimedia uses with Quartz, picture and movie previewing, etc. BTW, a whole computer is multimedia, so there was really no point in listing the apps. So the "multimedia-only" Cell processor isn't a possibility in a Mac? :confused:

What I find kind of tiring is that new technology processors are needed now in Apple's line-up and we're probably not gonna find out until the last minute as always. *sigh*

[wdlove]

What I find kind of tiring is that new technology processors are needed now in Apple's line-up and we're probably not gonna find out until the last minute as always. *sigh*

Not everything comes out at the last minute, remember the Mac mini. Sadly it seems that the lawsuit against Think Secret has put a damper on information. This information could be very critical.

[Peace]

Not everything comes out at the last minute, remember the Mac mini. Sadly it seems that the lawsuit against Think Secret has put a damper on information. This information could be very critical.

Post of the day candidate!
:)

Folks are saying nuthin new coming down the pipeline because Think Secret hasn't posted anything.
They can't folks..They're under gag orders..

[Lord Kythe]

On the one hand you state the CELL is a gaming-console costumized chip, therefore it cannot be used in a Macintosh, on the other hand you point to the capabilities of the 'PPC' since it is being used in gaming-consoles. Not only is that a strange argument, but I would recommend to read some more on the CELL and PPC before making such statements, as the CELL is a customized PPC just like the XBOX360 processor but one with additional potential beyond the gaming console.

IBM's POWER processors are not the same as their PPC chips. The Cell is based on IBM's POWER architecture (plus Sony's and Toshiba's contributions which are not present is PPC chips), NOT PPC. The G5 is based on a POWER derivative, but it is a PPC (AltiVec-optimized). The POWER architecture allows for great multi-threading/tasking (IBM uses them in servers), thus the Cell is using as many execution units as possible to compute a lot of operations simultaneously, while the PPC is efficient with multi-tasking, but is mainly efficient at crunching one operation at a time as fast as possible; thus its application in Desktop environments (multiple G5 cores makes a PM powerful in both areas, allowing very efficient xServes as well... I have a G4 xServe as my main server, can't wait for updates on G5 xserves!).

Nowhere did I mention the Cell was unadaptable to a PC environment, in fact I stated it would add tremendous pressure on developers (which means I believe it could be done).

I didn't point the capabilities, I pointed the popularity. Since there will be more orders for PPC chips, IBM won't have any choice but increase PPC production, which directly impacts us as Mac users. I don't know what good would come from IBM producing a highly specialized/customized POWER derivative chip for PPCs. That's all I meant by that.

Of course, any money injected into IBM can have positive impacts for the G5/successor, but, depending on the $'s source, it could bite use back in the @$$.

I'm not a computer engineer, like most of people who post here I read articles and use a bit of reasoning, I'm just trying to illustrate the unlikelyness of a Cell/derivative processor/co-processor in a Power Mac any time soon, IMHO. If Apple could find a way to easily add native support for a processor like the Cell, more power to them, but how likely is that compared to the 970MP or POWER5 version of a new PPC chip? I wonder...

[tdewey]

I read that the CELL is severely limited in that you can only link up two of them for multi-processing. Doesn't sound very promising to me.

Even if you discount the lack of 4-way multiprocessing (which isn't on the desktop right now) I'm sure that PPC and GPU improvements will outpace the CELL and its GPU within a year.

[Mitthrawnuruodo]

I read that the CELL is severely limited in that you can only link up two of them for multi-processing. Doesn't sound very promising to me.

Even if you discount the lack of 4-way multiprocessing (which isn't on the desktop right now) I'm sure that PPC and GPU improvements will outpace the CELL and its GPU within a year.Where do you get that from...???

From the above mentioned press release (http://www.sony.net/SonyInfo/News/Press/200411/04-1129BE/): Cell is a multicore chip comprising a 64-bit Power processor core and multiple synergistic processor cores capable of massive floating point processing, optimized for compute-intensive workloads and broadband rich media applications, including computer entertainment, movies and other forms of digital content.

[isgoed]

Apple have moved graphics over to core video/graphics - and core data etc for the the other elements. In effect they are abstracting the the system from the hardware - for Apple to shift over to something else eg Cell, it would appear they would only have to write code inside the system. Apps would be able to take advantage of the power afforded by the extra chips by default.

This is also similar to AltiVec (back in the day) - PS runs without it, but slower. Similar to 'ripples' working or not in Dashboard today with different GPUs.

The Cell chips are designed to work together. Higher end Macs could have lots and lower end could have fewer. This way apps could take advantage (a la AltiVec) and for the high end run faster with the greater workloads imposed on them.This system would indeed be feasable for a solution for getting a cell in a mac. The cell in this way acts like a kind of PCI or videocard with a coprocessor. However this is not how the cell works. In the cell the applications run directly on the cell. We will NOT see this in macs, for the following reasons:

Only one application will be able to use the cell (efficiently) Since running multiple apps would require subsequent clearance of the cell's cache and litterly bringing performance to an halt. (So solution is to have two cells if you want to run photshop and H.264 export at the same time)
It requires a rewrite of the OS to allow threads to run on a co-processor. I think this is not the idea behind cell so it will not even be possible.
Recompile the apps for the cell. This means that for one application you need fat binaries in order to make them compatible with older G4 and G5 systems. Taking in mind that it took apple a year to make fat binaries for the G5-64bit (which still are inferior to the 32 bit binaries, since they can't link to the 32 bit Carbon/Cacao/GUI/Aqua-Libraries/Frameworks) This will not happen within a year.
Developers not only need to compile for the cell but they probably also need to program with the cell's architecture in mind. Since no developer has done that, you will not see any benefits from a compiled cell application since it most likely only runs on the General Purpose CPU leaving all the other units running idle since the code never ever states it wants to utilize some unit for a secondary process.
So the future I see it, will be in the form of massive multithreaded CPU's in combination with killer altivec. That's what the Mach-O kernel is best at, so that is what will happen.

Even if there will be some way to fit a cell in a mac this will not constitute for a cell computer since the definition of a cell computer is something quite different than just offloading instructions. Similary that having a CPU and GPU does not constitute for a cell.

[thedoc1111]

So why can't the "primitive" PPE unit found in the current CELL processor of PS3 be replaced with a 970fx or power5 derivative?

Because then we would be back where we started - a high cost high transistor part with what is basically multiple supercharged Altivec units which do pretty much nothing to unoptimised code.


And also why can't a CELL act as a coprocessor for multimedia intensive apps?

And this is what I said all along.

I wish Apple would ship computers with an accessory Cell processor to handle CPU intensive repetitive tasks such as H.264 decoding/encoding. We'll have to wait and see, though.

[Neuro]

I find it 'funny' that Apple buys into IBMs new chips when the rest of the world is still screaming Intel! Intel! But, now everyone wants to go PowerPC and Apple is suddenly last in line for the same IBM goodies...

I can't help but think Apple would have kicked IBM into touch if they'd left Apple with the dregs of their R&D (or at least have some kind of priority supplier contract). It's almost inevitable that Apple will have a new next-gen chip from IBM very soon.

[oskar]

Not everything comes out at the last minute, remember the Mac mini. Sadly it seems that the lawsuit against Think Secret has put a damper on information. This information could be very critical.

Yeah, I was exaggerating :)
What I meant is that Apple never let's us know or at least never wants to let us know about anything until the day when the product will be officially announced.
Microsoft is telling us what processors the xbox will have months before it is released. The same goes for Sony with the PS3, which is supposed to include the Cell processor. Game consoles are a little off-topic but I mean that they give a hint about what processing the systems (still many months away) will have. It gives you something to look forward to.
Just what I was thinking... before people lecture me on Apple's secrecy. I know... I know...

[thedoc1111]

I can't help but think Apple would have kicked IBM into touch if they'd left Apple with the dregs of their R&D (or at least have some kind of priority supplier contract). It's almost inevitable that Apple will have a new next-gen chip from IBM very soon.

I think Apple's issue at the moment is that they are designing a major new motherboard design in conjunction with IBM and they are being more careful this time. The last 3 'generations' of G5 Mobos have been almost exactly the same (with the exception of the Single 1.8GHz model probably based on the iMac) even down to the same I/O problems with the FireWire 800 chip.

This new motherboard will have to be cheaper than the current design so Apple can stay competitive, be highly scalable as Apple cannot afford to go through the same design process in a year, and cannot afford to have the same problems as the previous board. It will also need to include many more forward-reaching technologies such as PCI Express (probably replacing PCI-X totally), DDR2, SATA-2 and hopefully wireless USB or Firewire if we are lucky. Support for a coprocessor would be a bonus...

[thedoc1111]

Only one application will be able to use the cell (efficiently) Since running multiple apps would require subsequent clearance of the cell's cache and litterly bringing performance to an halt. (So solution is to have two cells if you want to run photshop and H.264 export at the same time)
It requires a rewrite of the OS to allow threads to run on a co-processor. I think this is not the idea behind cell so it will not even be possible.
Recompile the apps for the cell. This means that for one application you need fat binaries in order to make them compatible with older G4 and G5 systems. Taking in mind that it took apple a year to make fat binaries for the G5-64bit (which still are inferior to the 32 bit binaries, since they can't link to the 32 bit Carbon/Cacao/GUI/Aqua-Libraries/Frameworks) This will not happen within a year.
Developers not only need to compile for the cell but they probably also need to program with the cell's architecture in mind. Since no developer has done that, you will not see any benefits from a compiled cell application since it most likely only runs on the General Purpose CPU leaving all the other units running idle since the code never ever states it wants to utilize some unit for a secondary process.


Fortunately, none of these reasons remove the possibility of the Cell being used as a coprocessor for Digital Video Encoding/Decoding, especially for the G5. Generally, a computer will only be running one render/movie compositing/compressing job at one time, and multiple Cells can be easily linked together to allow multiple simultaneous thread execution. Final Cut Pro gets 2.5x faster for $500 - Given what you pay for the software, I'd be happy!

[isgoed]

Fortunately, none of these reasons remove the possibility of the Cell being used as a coprocessor for Digital Video Encoding/Decoding, especially for the G5. Generally, a computer will only be running one render/movie compositing/compressing job at one time, and multiple Cells can be easily linked together to allow multiple simultaneous thread execution. Final Cut Pro gets 2.5x faster for $500 - Given what you pay for the software, I'd be happy!Yeah,

and fortunately this is already possible. Cinema 4D can use the graphics card for hardware rendering. You can buy MPEG encoding and decoding video cards, Sound cards do this. nVidia even wrote a library (for windows) that applications can leverage as to use the videocard as a co-processor, resulting in the equivalent of a 10GHZ Pentium 4! :eek:

So what's new?

[Xtremehkr]

I am sure that all of this development will help in some common areas. I see why it is hard to compete with PS3 and Xbox 360 though. Those platforms are going to move 10s of millions of units per year.

[thedoc1111]

So what's new?

Well, for a start, Apple is good at taking advantage of increases in hardware performance.

Does Avid use nVidia's extra movie processing power? No. Then the circuitry's existence is little more than proof of concept (especially in that it failed to work at all on MPEG-4 in the 6800 due to a mistake in the silicon)

Also, ATI doesn't do this, and the video card market is such that if they did, they would probably make the systems incompatible. Finally, the Video Processing Unit is much more specialised to the cell - burned in to accelerate specific codecs. The CELL is at least general purpose enough to accelerate any codec that optimised code is written for.

[GFLPraxis]

That would be interesting. I wonder how feasable that would be. Dual Processor G5 PowerMac with an additional cell processor for use with FCP and Logic.. Perhaps an add in PCI card with two cells for offloading. Would that even be possible? I suppose you'd have to rewrite your applications to utilize this applicaiton, but that would pretty slick.

Very feasible. The Cell core in the PS3 is capable of decoding 12 HD streams at ONCE. They demoed this on stage at E3.

It would suck for OS X, but can you imagine say a Cell processor with a FULL 2.5 GHz G5 in it? So for normal tasks it runs just as fast as a G5, but for multimedia you get the Cell speed boost. Of course it would cost a few hundred extra, but it would be cool. Say, add a $3500 PowerMac model with dual 2.7 GHz Cell G5's instead of just dual 2.7 GHz G5's.

[GFLPraxis]

Or not... considering sony saw it fit to include a GPU as well in their PS3 outside of the Cell processor. Cell is capable of general purpose computing, and IBM even goes and brags that they have used multiple operating systems on Cell. Heck, both sony and IBM do. At the Game Developers Conference, it was something they kind of highlighted on with 2 of their lectures that I had attended. The Cell processor is a capable processor. It is also a very powerful processor, and nowhere near what I would consider comparing to a GPU. I would compare it more to a dual core Opteron running at 3.2GHz... only this Opteron seems to have 8 additional AthlonXP processors around it(The SPEs) running at the same clock speed, all with their own cache and DMA channels, while at the same time, be able to punch the ammount of memory through per second that only 4 to 5 Opterons operating in NUMA would be able to do... in one chip.

Different purposes. He wasn't saying that the Cell WAS a GPU, of course Sony had to include a GPU in it. Thing is, Cell is extremely good at floating point calculations. So its very good for physics and HD decoding, which is precisely what Sony wants to use it for.

While it would cost a lot, the Cell is scalable, meaning that you could put a normal G5 processor in the Cell instead of the custom PowerPC core Sony is using and it WOULD be suitable for general computing. However, this would probably raise the cost of the processor by a couple hundred dollars so only in Apple's high end machines would it be used.

[rickvanr]

I don't know if this has been mentioned in the thread, because I only quickly skimmed through it, but with the cell being a new processor, with new architecture make it just as likely for apple to adopt this technology as it would be for them to throw an x86 processor (Intel, AMD). Unless the cell uses PPC, I don't think it will go into a mac.

[fBaran]

G6

First to say it.

G6 Powerbook!

[tdewey]

Where do you get that from...???

From the above mentioned press release (http://www.sony.net/SonyInfo/News/Press/200411/04-1129BE/):

From the EEtimes article page 2:

The high-end supercomputing option that Cell members have discussed may face hurdles as well. As currently architected, only two Cell processors can be directly attached to each other. A separate switch is needed to link more processors into the kinds of large arrays of CPUs used in supercomputers like IBM's BlueGene/L.

[tdewey]

Fortunately, none of these reasons remove the possibility of the Cell being used as a coprocessor for Digital Video Encoding/Decoding, especially for the G5. Generally, a computer will only be running one render/movie compositing/compressing job at one time, and multiple Cells can be easily linked together to allow multiple simultaneous thread execution. Final Cut Pro gets 2.5x faster for $500 - Given what you pay for the software, I'd be happy!

As noted above, according to the eeTimes multiple CELLS can not be easily linked together. Only 2 (though, of course, that may be enough).

[Lurch_Mojoff]

As much as I love reading Macrumors' forums, every time I go through a page one rumor therad I feel like I'm the only one that didn't get high on crack.

First, please stop with the G5 Powerbook! This is simply not going to happen. Just as Intel are having two different lines of procesors for desktops and laptops (namely Pentium 4 and Pentium M), so are Apple. Pentium M is just about as different from Pentium 4 as G4 is from G5. Why would you need G5 on a laptop anyway. Noone ever clamed Powerbooks are desktop replacement portable computers, they are simply high-end laptops. If the day, when POWER5 and Itanium (or at least Xeon) laptops hit the market, comes feel free to start again with this nonsence. But untill then, please, lets just call it NextGen Powerbook, or something.

Second, please understand that PowerPC architecture doesn't mean the same processor in different packaging. The chips in next-gen consoles are NOT G5, the only common thing is that they are both following the PPC guidelines. These chips are much simpler (and/or smaller) and therefore can run on higher clock speeds. IBM are not screwing anybody! The 970GX as fast as the design allows. Two years ago they (and Stevo) promised 3GHz, probably expecting to go to 90nm and 65nm manufacturing faster and so getting "free" MHz out of 970GX, but they (as well as Intel and AMD) hit the wall. Now there are two choices - either redesign of the G5 (970FX/970MP anyone) or hanging around till IBM can make 65nm SSOI 970GX' (if ever). For now it seems Apple are doing the second, while we are all hoping for the former (it all depends on Apple's plans for adopting Cell CPU - since a redesign costs money, if Apple are to use anyday Cell CPUs they may just decide to hold on for a while).

And third, now that I mentioned Cell, please understand that Cell is an architecture, a platform, not a single processor model. You would feel kinda' stupid if you say PowerPC is not a very good general purpose chip, won't you? Which PowerPC - the G5, the G4, the one in XBox360, one of the gazillion comunications equipment embeded chips form both Freescale and IBM?
But anyway, even the current version of the Cell is perfectly good CPU (it IS PS3's CPU after all). It has it's own downsides (e.g. inorder execution) but they are completely belied by the higher clock speed and the massive vectorised calculations performance (in both integer and fp). And I haven't heard anyone complaining that they can't reach 10000 characters per minute in Word, most complaints are of choppy h264 decoding, unsatisfactory game performance, rendering or encoding that takes forever - just the stuff that will benefit from the sinergistic processors in the Cell.
Yesterday a ton of people were vowing that PPC -> x86 transition will be a snap, and yet today a PPC -> PPC transition seems impossible? (On a sidenote, as wild as it may sound to some of you, judging by your posts, POWER is IBM's improvisation (and improvement if I may add) on PowerPC architecture; G5 is a POWER4 derivative and so is Cell's main core.)

To conclude I'll just add that IMHO a Cell workstation from Apple is not to be seen in the next 9 to 12 months, at the least. I'd rather expect a 970 redesign or even a POWER5 derivative, but again in order to cut R&D costs Apple may decide to just stick with the good old GX for another year. I think Cell's adoption by Apple will depend on the number of developers that are going to use Quartz2D Extreme, CoreImage, CoreVideo, etc., for these are things that Apple themselves can adjust to take advantage of the new architecture, therefore ensuring (relatively) painless early adoption.

Sorry for the rant, but I just don't want to see pointless, ala Slasdot, babbling here.

[Dr.Gargoyle]

First to say it.

G6 Powerbook!
First to say it. PLEASE STOP...just stop :rolleyes:

[tdewey]

Now there are two choices - either redesign of the G5 (970FX/970MP anyone) or hanging around till IBM can make 65nm SSOI 970GX' (if ever). For now it seems Apple are doing the second, while we are all hoping for the former (it all depends on Apple's plans for adopting Cell CPU - since a redesign costs money, if Apple are to use anyday Cell CPUs they may just decide to hold on for a while)

Dual-core baby! All the way. I'm waiting for the dual-core announcements at WWDC.

[thedoc1111]

As noted above, according to the eeTimes multiple CELLS can not be easily linked together. Only 2 (though, of course, that may be enough).

You are right - sorry for the misleading info there. Still, future iterations of the Cell should remove this (hopefully), and even so, 2 cells should be enough for anyone :)

[Lacero]

What else is out there anyway? The main contenders are PPC and Intel. Both Intel and AMD produce x86/64 chips only. Sure everyone also produces arm/strongarm chips, but theyre still weak, the fastest strongarm from Intel is used on higher end PDAs.

Whats left is MIPS, Ultrasparc, PA-RISC, Alpha and special purpose FPGA chips.

MIPS is dead. SGI was producing servers on Itanium which also died.

[GregA]

That would be interesting. I wonder how feasable that would be. Dual Processor G5 PowerMac with an additional cell processor for use with FCP and Logic.. Perhaps an add in PCI card with two cells for offloading. Would that even be possible? I suppose you'd have to rewrite your applications to utilize this applicaiton, but that would pretty slick.I assume the PS3 has ethernet and/or wireless networking. I wonder if Apple could write a "game" for the PS3 that turns it into a video compressing device running on Xgrid - so any Mac running DVDSP or iDVD could share the compression work with their PS3 games console.

To those thinking the Cell is not suitable for "normal" computing: IBM and Sony had a working prototype Cell-powered workstation back in November 2004. <snip> and another advantage is that the Cell seems to scale better than the Power4 (which the G5 is based on) so maybe we'll see 64 bit portable Macs not too far into the future. :DDid you read the article quoted? IBM haven't gotten Linux running yet which may indicate it's not as easy, and they also say that there are some significant programming challenges. In addition, they say the Cell chip will not be portable, it simply uses too much power - so it's not going into an Apple Powerbook.

[stephenli]

Anyone who thinks the cell will be a G6 is completely wrong - it is like saying that because my NVidia 6800 is extremely powerful, I should replace my CPU with one of them.


u should replace your CPU with a PS3 GPU - sony said that the performance doubled 6800

[mccoma]

What else is out there anyway? The main contenders are PPC and Intel. Both Intel and AMD produce x86/64 chips only. Sure everyone also produces arm/strongarm chips, but theyre still weak, the fastest strongarm from Intel is used on higher end PDAs.

Whats left is MIPS, Ultrasparc, PA-RISC, Alpha and special purpose FPGA chips.

MIPS is dead. SGI was producing servers on Itanium which also died.

to add to your point:
PA-RISC is dead, HP is moving to the Itanium :confused:
Alpha died long ago :mad:

I really can't imagine any switch to another 32-bit processor, G5 is 64-bits so any replacement should be also (so no Pentium-M for the Intel lovers).

[mccoma]

on the bad front, the CELL (in its current incarnation) is a big chip and will not be a nice thing to have in a portable (check the heat / wattage figures in some of the technical articles). As others have stated, dual is the current limit without a custom switch (well, I guess there are no quad macs currently). Also, check the memory you need to use with this thing, not very cheap.

Ars Technica is running an article (http://arstechnica.com/articles/paedia/cpu/xbox360-1.ars) about the XBox's new PPC. It has some nice stuff and might be a little closer to what the Mac would need.

I do wish Freescale would get their e700-base CPUs out and they would be low power. Always love to see "Future products in plan" on the roadmap.

[AidenShaw]

I really can't imagine any switch to another 32-bit processor, G5 is 64-bits so any replacement should be also (so no Pentium-M for the Intel lovers).

What on earth makes you think that you need more than 4 GiB of RAM on a laptop?

Even though the PPC970 does support 64-bits, Apple is still stuck in 32-bit land! 10.3 is pure 32-bit addressing. 10.4 has some 64-bit stuff - *unless* you want to use graphics or a GUI or Carbon or Cocoa - that's still 32-bit only.

And for "switching to another 32-bit processor" - did you not notice that both Intel and AMD are all 64-bit now (on the higher end)? It should be obvious that Apple would choose an x64 architecture to jump to....

[alandail]

instead of going into the high end powermacs, doesn't it make a ton of sense for the cell to be in the next Mac Mini? Imagine a Mac mini with the following

Cell CPU
Blu-Ray Drive
HDMI out
Digital optical out
HDTV input
Tivo like software

Who here wouldn't buy one of those for each of their HDTVs? I'd order 3 the day they were announced. It would play DVDs, blu ray DVDs, CDs, iTunes, play quicktime, video games, browse the web, etc, etc, etc. The G4 in the Mac mini doesn't have the horsepower to do all of that. Shoot, it takes dual 2GHz G5s to decode 1080i video. The cell can do all of that. Tiger should run just fine on it with the PowerPC and Apple already has system libraries already in place that seem well suited to the cell.

[GFLPraxis]

I don't know if this has been mentioned in the thread, because I only quickly skimmed through it, but with the cell being a new processor, with new architecture make it just as likely for apple to adopt this technology as it would be for them to throw an x86 processor (Intel, AMD). Unless the cell uses PPC, I don't think it will go into a mac.

The Cell is not a new architecture. It is a PowerPC processor with secondary 'cells' that are designed to do floating point calculations attached.

[mccoma]

What on earth makes you think that you need more than 4 GiB of RAM on a laptop?

Video - if I need all that speed past a G4 on my laptop, then I would imagine it is a mobile Final Cut Pro station and video takes a lot of RAM.

I would also imagine, if Apple is switching processors, then the whole line is going to go (no sense in doing it by half - that would just confuse buyers).

Even though the PPC970 does support 64-bits, Apple is still stuck in 32-bit land! 10.3 is pure 32-bit addressing. 10.4 has some 64-bit stuff - *unless* you want to use graphics or a GUI or Carbon or Cocoa - that's still 32-bit only.

Well, I don't care about < 10.4 (as they would not port the older OS to the new platform). Having a GUI does not exclude having parts of your app be 64-bits - to quote Apple (http://developer.apple.com/macosx/64bit.html) :

It is important to note that in Tiger, the support for 64-bit programming does not extend throughout the entire set of APIs available on Mac OS X. Most notably, the Cocoa and Carbon GUI application frameworks are not ready for 64-bit programming. In practical terms, this means that the "heavy lifting" of an application that needs 64-bit support can be done by a background process which communicates with a front-end 32-bit GUI process via a variety of mechanisms including IPC and shared memory.

Takes a little thought as a developer, but not something too far out.

And for "switching to another 32-bit processor" - did you not notice that both Intel and AMD are all 64-bit now (on the higher end)? It should be obvious that Apple would choose an x64 architecture to jump to....

I specifically mentioned a 32-bit member of the Intel line, that is being mentioned heavily in another (non-CELL) thread. I did not say anything about the 64-bit x86. I assume they would not switch to another 32-bit processor, the 64-bit x86's are not 32-bit processors.

Looking at the 64-bit x86's, I am having a tough time believing a "real - decent battery life" portable could be built around those processors.

I still think the 970 is a fine 64-bit processor, my only holdup on the PowerMac is the AGP bus. I will buy a PowerMac when AGP is replaced with PCI-Express.

[hcorf]

while we're on the g4/g5 powerbook conversation...

how much extra performance would you be likely to see in a g5 powerbook, it would still have to have a slower hdd and a lower fsb.

my point is, how does a top end 1.67ghz powerbook compare to a 1.6ghz imacg5?

i am sure a lot of the performance of the g5 comes from apple removing all the bottlenecks to performance, which exist in so many pc's these days. I think there would be a lot of bottlenecks in a laptop.

[isgoed]

The Cell is not a new architecture. It is a PowerPC processor with secondary 'cells' that are designed to do floating point calculations attached.I think it is a new architecture. The cell uses PowerPC instructions because that was just convenient for IBM. But that doesn't mean that it is PowerPC compliant.

You can for example say that the G4 is a powerPC core + additional altivec instructions. This is simply proved by the fact that you can compile G3 code and it will run on the G4.

And with that same argument I highly doubt that G3 code will run on a cell. Simply because the threads that run on cell are not regular threads and the task of the general Purpose processor is to be the arbiter over the cells. It doesn't understand doing things itself. Furthermore the way the cell works requires a whole lot more cell specific instructions and communications commands in the cell. I think this is even more illustrated by the comment above that Linux does not run on Cell. If Cell was just general PPC this would be a piece of cake.

[stephenli]

so, would it be closer to expect XBox360's CPU instead of Cell?
M$ use G5s to run XBox360 & develop its game
PowerPC Based 3.2Ghz multi-core right?

As i remembered, PS2's architecture is linux based, so as PS3?! I am not sure. But Cell is originally developed for Cluster computing, and this is also somehow close to sth like XGrid with G5s....so it may also be possible.

Question: what is faster? and which generates more heat?
Cell@3.2Ghz or XBox360 CPU@3.2Ghz or current G5 ?

[840quadra]

It is funny to watch these threads. As someone else said, we don't know what is coming from Apple in the near future, and the same goes for IBM. What I am talking about, is the next advance in the Macintosh line.

Apple has an extremely limited Market share in the Home and business PC market, yet they have been consistent in having some of the fastest computers on the market. So the Powermac isn't the fastest home / office system on the market now, oh well, it doesn’t necessarily mean that Apple is going to stand by and stop working to produce products in the top 10 or 20% of fastest consumer and professional computers. The only bad thing about the current situation is that it opens the doors more for the Apple nay-Sayers to bop in and slam the Macintosh and Powerbook line.

Are the computers we own and buy THAT slow? Heck no, they are not the fastest in the market, but they still have allot going for them. I am sure that Apple will take care of us, and with the IBM leak (a while back) in regards to Dual Core G5 Processors (that seems to have been forgotten on recent debates) may come to market sooner then later.

On the Intel subject, I have yet to see an official announcement from Apple or Intel, that the next Powermac will be X86, or even have an Intel core. I don’t think anyone should commit suicide, or loose sleep over this at this point.

I am not, and will not be in the market for a faster computer in the near future, but that doesn’t keep me from being optimistic that Apple may soon release more then a speed bumped upgrade to the Powermac. Don't let the recent 2.7 GHZ speed bump make you think a larger upgrade isn't coming soon. If you look at the history of the Macintosh line, speed bumps sometimes only lasted a couple of months, before the release of an other bump, and or replacement model with much faster gains or new Technology.

I know this is Macrumors, but a BIG part of Apple computer in the last 5 years, as been their love for the element of surprise!

just my thoughts, nothing more.

840

[Platform]

Why in the ---- would Apple not make use of this processor, assuming it's all that it's cracked up to be?

And why in the ---- is Apple putting along at 2.something Ghz when the new XBOX has an IBM PowerPC-based CPU with 3 symmetrical cores running at 3.2GHz each.

:mad: :mad: :mad:

WTF?!

Please...NO MORE OF THESE POSTS...it has been explained so many times ;)

[AidenShaw]

Looking at the 64-bit x86's, I am having a tough time believing a "real - decent battery life" portable could be built around those processors.

That's why Intel has the Napa64 and Meram projects underway - to move the Pentium M line to x64....

[alandail]

the cell is a powerpc - there should be no reason why existing powerpc code wouldn't run unmodified on the cell - only it wouldn't take advantage of the cells themselves. But Apple building tiger for the cell could have appropriate OS stuff use the cell - stuff like quarts extreme, core image, quicktime codecs, etc, thus speed up many apps without the apps themselves needing to be rewritten.

[minimax]

IBM's POWER processors are not the same as their PPC chips. The Cell is based on IBM's POWER architecture (plus Sony's and Toshiba's contributions which are not present is PPC chips), NOT PPC. The G5 is based on a POWER derivative, but it is a PPC (AltiVec-optimized). The POWER architecture allows for great multi-threading/tasking (IBM uses them in servers), thus the Cell is using as many execution units as possible to compute a lot of operations simultaneously, while the PPC is efficient with multi-tasking, but is mainly efficient at crunching one operation at a time as fast as possible; thus its application in Desktop environments (multiple G5 cores makes a PM powerful in both areas, allowing very efficient xServes as well... I have a G4 xServe as my main server, can't wait for updates on G5 xserves!).

Nowhere did I mention the Cell was unadaptable to a PC environment, in fact I stated it would add tremendous pressure on developers (which means I believe it could be done).

I didn't point the capabilities, I pointed the popularity. Since there will be more orders for PPC chips, IBM won't have any choice but increase PPC production, which directly impacts us as Mac users. I don't know what good would come from IBM producing a highly specialized/customized POWER derivative chip for PPCs. That's all I meant by that.

Of course, any money injected into IBM can have positive impacts for the G5/successor, but, depending on the $'s source, it could bite use back in the @$$.

I'm not a computer engineer, like most of people who post here I read articles and use a bit of reasoning, I'm just trying to illustrate the unlikelyness of a Cell/derivative processor/co-processor in a Power Mac any time soon, IMHO. If Apple could find a way to easily add native support for a processor like the Cell, more power to them, but how likely is that compared to the 970MP or POWER5 version of a new PPC chip? I wonder...


wow, great response, and thanks for explaining :o

[Chryx]

To those of you stating that Cell is a multimedia chip and not suitable for standard processing...


it has a 64bit PowerPC core with full VMX/Altivec support, this is augmented by an array of independent vector units with their own memory.

a 4Ghz+ Cell would quite happily run OSX, and very quickly at that, obviously, the SPE's would need directly tuned code... which is where things like Coreimage come into the equation.

[poundsmack]

we dont know if it will be able to run OSX or not as it is. when it starts running linux then we will know if in its current state it is ready for a full featured OS.

[Chryx]

we dont know if it will be able to run OSX or not as it is. when it starts running linux then we will know if in its current state it is ready for a full featured OS.

the PPE featureset is lined up to run AS/400 according to IBM.

and if it'll run that, it most definately will run OSX (since AS/400 needs all the features OSX needs AND additional features that OSX doesn't)

[poundsmack]

interesting :)

[shamino]

the PPE featureset is lined up to run AS/400 according to IBM.

and if it'll run that, it most definately will run OSX (since AS/400 needs all the features OSX needs AND additional features that OSX doesn't)
It's my understanding that IBM uses various forms of the PowerPC (or POWER) for most of their systems - AS/400 minis, their UNIX servers (of various names) and their non-Windows desktop boxes.

The only big-iron architectecture that's not PPC/POWER based is their s/390 boxes. The 360/370/390 processor architecture is very very different from everything else, so systems based on it will probably never move to anything different. Porting the various operating systems (MVS/TSO, VM/CMS, etc.) would be virtually impossible, since the code is decades old and contains huge amounts of hand-optimized assembly language.

[sinisterdesign]

instead of going into the high end powermacs, doesn't it make a ton of sense for the cell to be in the next Mac Mini? Imagine a Mac mini with the following

Cell CPU
Blu-Ray Drive
HDMI out
Digital optical out
HDTV input
Tivo like software

if you tricked out a mini like that would it then be called a "maxi"??

[wdlove]

Dual-core baby! All the way. I'm waiting for the dual-core announcements at WWDC.

We definitely need something special to come out of WWDC. So many possibilities and nothing definite.

[Lacero]

if you tricked out a mini like that would it then be called a "maxi"??If the tablet rumors hold any weight, perhaps Apple will call it the Maxi iPad?

[271]

If the tablet rumors hold any weight, perhaps Apple will call it the Maxi iPad?


LOL!!!

[swissmann]

I'd like to see Apple do something amazing with their hardware so that when it is compared to anything else on the market it just blows it away and Intel couldn't do anything about it. I don't know where the answer lies but the G5 isn't even at 3 GHz yet, something better be done.

[AidenShaw]

It's my understanding that IBM uses various forms of the PowerPC (or POWER) for most of their systems - AS/400 minis, their UNIX servers (of various names) and their non-Windows desktop boxes.

The only big-iron architectecture that's not PPC/POWER based is their s/390 boxes.

Your "understanding" would be a lot better if you simply looked at IBM's product website.

You'll have to look a *long* time to find the one and only product that IBM sells with a PPC processor. Except for the zSeries (new name for s/390), every product except for one is either POWER or x86 based.

IBM is not a PPC vendor. The only PPC they offer is one system with a PPC970 (which, of course, isn't as much a real PPC as it is a cut-down POWER4 chip).

[wizard]

You will be happy to note that patches are already being submitted to the Linux kernel to support new IBM hardware. Take a lokk at the development logs and some of the on line resources.

Dave


the PPE featureset is lined up to run AS/400 according to IBM.

and if it'll run that, it most definately will run OSX (since AS/400 needs all the features OSX needs AND additional features that OSX doesn't)

[techgeek]

Looks like they got Linux up and running.
http://techon.nikkeibp.co.jp/english/NEWS_EN/20050525/105050/
Mac OS X next? ;)

[Lurch_Mojoff]

There you go - the next Powermac (http://techon.nikkeibp.co.jp/english/NEWS_EN/20050525/105050/)!!!

:p :p :p

EDIT: ooops... techgeek beat me to it! Cheers!

[Lacero]

Could this be the prototype motherboard (http://techon.nikkeibp.co.jp/english/NEWS_EN/20050525/105050/?SS=imgview_e&FD=-1016875769) for the next PowerMac G6?

[shamino]

Your "understanding" would be a lot better if you simply looked at IBM's product website.

You'll have to look a *long* time to find the one and only product that IBM sells with a PPC processor. Except for the zSeries (new name for s/390), every product except for one is either POWER or x86 based.
PPC and POWER are the same architecture. And I made a point of mentioning both. You ask me to look through IBM's products (which I have done on many occasions) but you obviously didn't bother reading the message you replied to.

As for x86, those are all PCs. (A PC-server with a dozen hard drives in it is still a PC, no matter what the marketing department wants to call it.)

[AidenShaw]

PPC and POWER are the same architecture. And I made a point of mentioning both. You ask me to look through IBM's products (which I have done on many occasions) but you obviously didn't bother reading the message you replied to.

As for x86, those are all PCs. (A PC-server with a dozen hard drives in it is still a PC, no matter what the marketing department wants to call it.)

http://forums.macrumors.com/showpost.php?p=1484851&postcount=77
http://forums.macrumors.com/showpost.php?p=1484851&postcount=78

How about a x64 server with a dozen CPUs in it - still a PC?

[wdlove]

Could this be the prototype motherboard (http://techon.nikkeibp.co.jp/english/NEWS_EN/20050525/105050/?SS=imgview_e&FD=-1016875769) for the next PowerMac G6?

Is this an actual working Cell processor? Where is it actually being used?

Or is this a fake prototype of what someone thinks it will look like?

[alandail]

Is this an actual working Cell processor? Where is it actually being used?

Or is this a fake prototype of what someone thinks it will look like?

It's the IBM two cell blade card that is discussed in the links a couple of posts ago (the two posts right before the one you quoted - the picture is from the article). Apparently despite what some on here assume, IBM thinks the cell will make for much more powerful servers than their current chips.

[Electric Monk]

Some information that may help you to better understand this whole Cell business.

PowerPC is an architecture just like x86. POWER and Cell are both PowerPC. The PPE is a brand new chip. It is not a POWER4 derivative like 970's (G5's)are. It does have some features from POWER5 like SMT though.

The Xbox 360's three core processor is in all probability a set of three PPE's. The PPE is a PowerPC processor probably based on some work IBM did back in 1999/2000 with the first processor to crack the 1 GHz barrier. i.e. it's simple, efficient at in order code and scales very high. IBM owns the PPE, when Microsoft wanted something they probably said, "Well we have this PPE thing that Sony-Toshiba are using for Cell's central processor, but we own the PPE itself and we're happy to customize it to your specs" Microsoft says ok, adds a custom VMX (The VMX-128, the regular VMX is also known as Altivec) and you have a deal. Despite the fact that Xbox 360 is using the PPE, the rest of their system has nothing to do with Cell.

That's because outside of the PPE the Cell is joint intellectual property (IP) of Sony-Toshiba-IBM (STI). The SPE's, the broadband bandwidth, etc. With additional stuff from other companies like Rambus's Flex/IO interface and they're ultra fast XDR (Extreme Data Rate) RAM.

The PPE even with 8 SPE's is not suitable for Macs. That's because while a PPE is roughly equivalent to a pair of of G4's (NOT G4+'s) at 1.4-1.6 GHz it's not good at a lot of things. The original G4 (7400 series) was a G3+Altivec and that's what a PPE can kinda be compared too, not that they ever got up to 1.4 GHz. Modern G4's are G4+'s which are quite different from G4's. The reason the PPE matches a pair of G4's is because of SMT (Simultaneous Multi Threading, like Intel's Hyperthreading but a better design). G3's and G4's are fairly in order processors so code written for them (like the vast majority of current Mac code) will run fairly well on a PPE. However newer code written for G4+'s and G5's will not run that well since the PPE is poor at out of order code, has poor branch prediction, and suffers a fairly large branch misdirection penalty. Therefore while it can be used, it wouldn't be the best solution in its current form. Luckily Cell is more of an architecture then a set of processors, and so the PPE can be modified, as can the number of SPE's.


Apple has a bit of a problem with upcoming processors. The 970 series hasn't seen an improvement since the FX model, and both the GX and the MP are not on the scene yet. IBM seems to be unable/unwilling to make a low power laptop processor, at least until all process improvements are up and running at 90nm. While the G4+ 7448 chip and the e600 series are on the horizon, Freescale is half a generation behind IBM in manufacturing, and may fall further behind. So short (7448) and medium (e600) laptop chips can come from them, but their track record needs watching and long term (e700) depends heavily on unknowns.

IBM's chip division uses a vast amount of the companies resources, but the return-on-investment sucks. Therefore they may have little to no interest in continuing to improve the 970's, or towards producing a POWER5 lite. Or Apple is unwilling to invest the money required for them to continue.

On the other hand Cell is going into probably a 100 million PS3's, and the PPE core will be in something like half that many Xbox 360's. Plus Nintendo is also using an IBM processor (could be a 970GX/MP if they come out, or a PPE, or a custom variant of either). Although Sony and Tosiba will be using their own fabs when they come online at 65nm, and Microsoft will be sourcing their PPE as soon as they find some 65nm fabs, IBM will probably be making a whole lot of Cell's/PPE's at 90nm and some at 65nm. Cash flow. Additionally if IBM also wants to win the processors for the next next gen systems (and I'm sure they do) they'll be investing lots of money in the Cell and PPE and derivatives while selling the improved models to as many people as possible to keep the cash coming in while they wait for the really big orders from the next next gen systems.

Therefore Apple jumping on a custom Cell will make IBM happy because it means moving upwards of 2 million Cell's a year for some time once all Apple systems are using variations of the Cell. Now say 6-10 million isn't a huge amount versus 200 million Sony/Microsoft/Maybe Nintendo but IBM won't be making most of those. Add in some workstation sales, and the money they've already made designing it for various customers and they should have plenty of money to keep working on the Cell/PPE for the next iteration of console wars around 2010.

Jumping on the Cell will make Apple happy because the Cell is going to have dozens of times the resources thrown at then the 970's or some Apple custom POWER5 lite (remember Apple is basically the only customer for 970's, IBM sells a handful in a blade server and that's it). I'm sure a beefed up PPE with more out of order and better branch prediction OR a few PPE's clustered together will make a fine future system once you toss in a few SPE's as well. Additionally Cell would scale very easily from just a single PPE with maybe 1 SPE in a Mac Mini or iBook, to multiple PPE's and SPE's in PowerMacs.


As to those pointing out Cell's size, yes it is big. But this is very first gen stuff at 90nm. Cell was designed to be built at 65nm but both Microsoft and Sony decided to kick off the next round before 65nm fabs are up and running so the first ~6 months for Microsoft and the first couple of months for Sony are stuck using 90nm with all the additional costs that implies.

At 65nm the Cell will be much more affordable, and the current die size should shrink with more engineering dollars. By late 2006/early 2007 Apple would probably be able to buy Cells running at 4+ GHz at 65nm with their own heavily beefed up PPE since IBM prides themselves at custom creations.

Although a 970MP or two would probably be able to match the performance at that point in time, going forward it would probably lose out real fast as it hits scaling limits (which the 970FX has run smack into right now) and the Cell climbs towards 6 GHz.

That gives Apple one to two years to begin rewriting/optimizing for in order code, and optimizing for those SPE's plus all the new internal bandwidth. Independent developer code won't see much of a hit or any depending on how beefed up the PPE(s) are and lots of apps will be able to benefit from the SPEs.


Anyway, that's my two cents.

[blitzkrieg79]

Some information that may help you to better understand this whole Cell business.

PowerPC is an architecture just like x86. POWER and Cell are both PowerPC. The PPE is a brand new chip. It is not a POWER4 derivative like 970's (G5's)are. It does have some features from POWER5 like SMT though.

The Xbox 360's three processors are in all probability three PPE's. The PPE is a PowerPC processor probably based on some work IBM did back in 1999/2000 with the first processor to crack the 1 GHz barrier. i.e. it's simple, efficient at in order code and scales very high. IBM owns the PPE, when Microsoft wanted something they probably said, "Well we have this PPE thing that Sony-Toshiba are using for Cell's central processor, but we own the PPE itself and we're happy to customize it to your exacting specs" Microsoft adds a custom VMX (The VMX-128, the regular VMX is also known as Altivec) and you have a deal. Despite the fact that Xbox 360 is using the PPE, the rest of their system has nothing to do with Cell.

That's because outside of the PPE the Cell is joint intellectual property (IP) of Sony-Toshiba-IBM (STI). The SPE's, the broadband bandwidth, etc. With additional stuff from other companies like Rambus's Flex/IO interface and they're ultra fast XDR (Extreme Data Rate) RAM.

The PPE even with 8 SPE's is not suitable for Macs. Thats because while a PPE is roughly equivalent to a pair of of G4's (NOT G4+'s) at 1.4-1.6 GHz. The original G4 (7400 series) was a G3+Altivec. Modern G4's are G4+'s which are quite different from G4's. The reason the PPE matches a pair of them is because of SMT (Simultaneous Multi Threading, like Intel's Hyperthreading but a better design). G3's and G4's are fairly in order processors so code written for them (like the vast majority of current Mac code) will run fairly well on a PPE. However newer code written for G4+'s and G5's will not run that well since the PPE is poor at out of order code, has poor branch prediction, and suffers a fairly large branch misdirection penalty.


Apple has a bit of a problem with upcoming processors. The 970 series hasn't seen an improvement since the FX model, and both the GX and the MP are not on the scene yet. IBM seems to be unable/unwilling to make a low power laptop processor, at least until all process improvements are up and running at 90nm.

IBM's chip division uses a vast amount of the companies resources, but the return-on-investment sucks. Therefore they may have little to no interest in continuing to improve the 970's, or towards producing a POWER5 lite. Or Apple is unwilling to invest the money required for them to continue.

On the other hand Cell is going into probably a 100 million PS3's, and the PPE core will be in something like half that many Xbox 360's. Plus Nintendo is also using an IBM processor (could be a 970GX/MP if they come out, or a PPE, or a custom variant of either). Although Sony and Tosiba will be using their own fabs when they come online at 65nm, and Microsoft will be sourcing their PPE as soon as they find some 65nm fabs, IBM will probably be making a whole lot of Cell's/PPE's at 90nm and some at 65nm. Cash flow. Additionally if IBM also wants to win the processors for the next next gen systems (and I'm sure they do) they'll be investing lots of money in the Cell and PPE and derivatives while selling the improved models to as many people as possible to keep the cash coming in while they wait for the really big orders from the next next gen systems.

Therefore Apple jumping on a custom Cell will make IBM happy because it means moving upwards of 2 million Cell's a year for some time once all Apple systems are using variations of the Cell. Now say 6-10 million isn't a huge amount versus 200 million Sony/Microsoft/Maybe Nintendo but IBM won't be making most of those. Add in some workstation sales, and the money they've already made designing it for various customers and they should have plenty of money to keep working on the Cell/PPE for the next iteration of console wars around 2010.

Jumping on the Cell will make Apple happy because the Cell is going to have dozens of times the resources thrown at then the 970's or some Apple custom POWER5 lite (remember Apple is basically the only customer for 970's, IBM sells a handful in a blade server and that's it). I'm sure a beefed up PPE with more out of order and better branch prediction OR a few PPE's clustered together will make a fine future system once you toss in a few SPE's as well.


As to those pointing out Cell's size, yes it is big. But this is very first gen stuff at 90nm. Cell was designed to be built at 65nm but both Microsoft and Sony decided to kick off the next round before 65nm fabs are up and running so the first ~6 months for Microsoft and the first couple of months for Sony are stuck using 90nm with all the additional costs that implies.

At 65nm the Cell will be much more affordable, and the current die size should shrink with more engineering dollars. By late 2006/early 2007 Apple would probably be able to buy Cells running at 4+ GHz at 65nm with their own heavily beefed up PPE since IBM prides themselves at custom creations.

Although a 970MP or two would probably be able to match the performance at that point in time, going forward it would probably lose out real fast as it hits scaling limits (which the 970FX has run smack into right now) and the Cell climbs towards 6 GHz.

That gives Apple one to two years to begin rewriting/optimizing for in order code, and optimizing for those SPE's plus all the new internal bandwidth. Independent developer code won't see much of a hit or any depending on how beefed up the PPE(s) are and lots of apps will be able to benefit from the SPEs.


Anyway, that's my two cents.

Yep I pretty much agree with you, why would IBM want to invest money in for example power5 lite version when the only customer will seem to be Apple. With Cell they have something new, something that it seems will be used by few top players in the tech game... SO obviously over the long run the research and further development of it will won't be as expensive as some exclusive Apple chip... People are saying that the CELL is not for everyday computig that it is specialized, they keep on forgetting that CELL is not one processor, its more of an idea, more of a broad architecture, CUSTOMIZABLE...

Anyway, I am more than sure they will be able to develop an Apple friendly verion of Cell-like processor in lets 12-18 months, they already got the basics and whose to say that Apple and IBM hasn't been working on such a design for the past 6-12 months...

Anyway, I do believe Apple in 1-2 years will use some derivative of the CELL design, IBM and Sony already declared they will develop graphic workstations running on CELLs so I am guessing that the CELL can be customized for more of everyday computing and not be as specialized as for the PS3 console...

[isgoed]

Some information that may help you to better understand this whole Cell business.You do not provide just information here. You speculate as well. It's kinda misleading.
PowerPC is an architecture just like x86. POWER and Cell are both PowerPC. The PPE is a brand new chip. It is not a POWER4 derivative like 970's (G5's)are. It does have some features from POWER5 like SMT though.Can you provide a link? I have seen nowhere that the cell can actually run PPC code.
IBM owns the PPE, when Microsoft wanted something they probably said, "Well we have this PPE thing that Sony-Toshiba are using for Cell's central processor, but we own the PPE itself and we're happy to customize it to your specs" Microsoft says ok, adds a custom VMX (The VMX-128, the regular VMX is also known as Altivec) and you have a deal. Despite the fact that Xbox 360 is using the PPE, the rest of their system has nothing to do with Cell.You are really making this up? are you?
Therefore Apple jumping on a custom Cell will make IBM happy because it means moving upwards of 2 million Cell's a year for some time once all Apple systems are using variations of the Cell. Now say 6-10 million isn't a huge amount versus 200 million Sony/Microsoft/Maybe Nintendo but IBM won't be making most of those. Add in some workstation sales, and the money they've already made designing it for various customers and they should have plenty of money to keep working on the Cell/PPE for the next iteration of console wars around 2010. Worldwide PC sales are only 43 Million (Q2 2004) (http://www.xbitlabs.com/news/other/display/20040722093125.html). So about 170 million a year. I highly doubt that console sales will top that (or even just cell sales). And I don't think IBM is bothered with the "small market" of Apple. I think more people buy Apple computers than IBM POWER computers.
Although a 970MP or two would probably be able to match the performance at that point in time, going forward it would probably lose out real fast as it hits scaling limits (which the 970FX has run smack into right now) and the Cell climbs towards 6 GHz.The initial rumours claimed cell speeds of 4.8Ghz. Now when the cell is here it only runs at 3.2Ghz. I think that as long as Mhz is concerned all processors face the same problems. That the 970 ran into its scaling limits is nonsense.
Anyway, that's my two cents.Glad you made that clear. You are up to date on computer developments, but I think you mix up fact and fiction too much. Putting cell into Apple requires a lot of hurdless to be overcome and certainly aint as easy as you make believe.

[Chryx]

[quote]Can you provide a link? I have seen nowhere that the cell can actually run PPC code.[quote]

Cell is a PowerPC processor with Altivec and an array of dedicated independent vector units.

http://www.pcstats.com/articleimages/200502/cell_3jp.jpg <- that thing labelled 'PPE' is a POWERPC/AS core.


As for the clockspeeds, the apparent reason they're clocking them at 3.2Ghz (on a 90nm process no less) is thermally motivated.

[isgoed]

Cell is a PowerPC processor with Altivec and an array of dedicated independent vector units. Well I made this comment before, but I'll make it again. We know that the PPE unit is based on the POWER architecture. This however does not mean that it can actually run PPC code. For example. The Athlon 64 is based on x86, BUT if they had not added extra instructions, the Athlon 64 could not run x86-32 bit. It all comes donw to the art of logic reasoning.

What is also interesting is the folowing: The Playstation 3 emulates PS1 and PS2 for backwards compatibility (http://games.slashdot.org/article.pl?sid=03/09/02/1517232&tid=203&tid=212&tid=10) Both the PS1 and PS2 are PPC computers. Of course it is not exactly clear what is being emulated. Maybe the code itself can run unemulated and just the hardware is emulated. Of notice is that also the PS2 emulates the PS1.
As for the clockspeeds, the apparent reason they're clocking them at 3.2Ghz (on a 90nm process no less) is thermally motivated.Naturally.

[gregoryp]

As mentioned earlier...
"Looks like they got Linux up and running.
http://techon.nikkeibp.co.jp/englis...0050525/105050/
Mac OS X next?"

OSX is Linux based. Is the jump to OSX as difficult as some have said ("...cell is not designed to run a full fledged operating system, etc.")? Does this answer the PPC compatible question?

[Tulse]

OSX is Linux based.

OS X is based on BSD UNIX, not Linux. And there is a vast difference between getting Linux (or even Darwin) running on a chip and getting the entire OS X experience ported (heck, Darwin can run on x86 chips, but that doesn't mean that OS X does).

[isgoed]

OSX is Linux based.Nooooo. The evil has fallen over us. Can we still be saved, or are we doomed by denying our heritage and diverting to the heartless anarchistic monster called Linux.

In short: OSX is UNIX based.Is the jump to OSX as difficult as some have said ("...cell is not designed to run a full fledged operating system, etc.")? Does this answer the PPC compatible question?I am confident that cell can even run the king of operating systems. However your argument that should support this is invalid. A: Linux can be trimmed down to the simplest of OS so being able to run it proves nothing. B No, it does not answer the PPC compatible question since Linux can run on every processor architecture.

[alandail]

OS X is based on BSD UNIX, not Linux. And there is a vast difference between getting Linux (or even Darwin) running on a chip and getting the entire OS X experience ported (heck, Darwin can run on x86 chips, but that doesn't mean that OS X does).

The cell is a PPC + other processors on a single chip - the PPC is the controller and should be able to run OS X unmodified. There is really no reason to think otherwise. In fact, there is every reason to think that the PPC that is inside the cell is the same PPC that MS is using in the new XBox. While the OS clearly could benefit from taking advantage of the other processors in the cell, those changes wouldn't have to be done to get the OS up and running.

And for the record, there is every reason to think the full OS X (minus classic) runs on x86 hardware in apple's labs. It would be a mistake for Apple to not maintain compatible builds of each major release.

[Chryx]

Well I made this comment before, but I'll make it again. We know that the PPE unit is based on the POWER architecture. This however does not mean that it can actually run PPC code. For example. The Athlon 64 is based on x86, BUT if they had not added extra instructions, the Athlon 64 could not run x86-32 bit. It all comes donw to the art of logic reasoning.

What is also interesting is the folowing: The Playstation 3 emulates PS1 and PS2 for backwards compatibility (http://games.slashdot.org/article.pl?sid=03/09/02/1517232&tid=203&tid=212&tid=10) Both the PS1 and PS2 are PPC computers. Of course it is not exactly clear what is being emulated. Maybe the code itself can run unemulated and just the hardware is emulated. Of notice is that also the PS2 emulates the PS1.
Naturally.

Firstly, the Athlon64 is an EXTENDED AthlonXP (wider registers, SSE2 support in microcode through the existing execution hardware, ondie memory controller, hypertransport for I/O), it doesn't HAVE '32bit added instructions', it's a 32bit core that's been 'stretched' for 64bit usage.

secondly, the Playstation 2 is MIPS based, not PowerPC, and PS2 uses the entirety of a PS1 as it's I/O controller, hence the backwards compatability.


using Logic and reasoning is all well and good, but you might want to check your reasoning using correct variables before implying everyone else is wrong.

[isgoed]

Firstly, the Athlon64 is an EXTENDED AthlonXP (wider registers, SSE2 support in microcode through the existing execution hardware, ondie memory controller, hypertransport for I/O), it doesn't HAVE '32bit added instructions', it's a 32bit core that's been 'stretched' for 64bit usage.Please!! Even without having read this I could deduce this. If you have only 64 bit adresses an array of 32 bit Ptr's does not work. Take a look at: http://arstechnica.com/cpu/03q1/x86-64/x86-64-4.html You can see that the Athlon 64 even has a whole logic-area full of transistors to support Legacy Mode. It all comes down to logic and a good memory (march 2003)
secondly, the Playstation 2 is MIPS based, not PowerPC, and PS2 uses the entirety of a PS1 as it's I/O controller, hence the backwards compatability.OK, the PS(1/2) is MIPS. I got confused with RISC, which they both are.

using Logic and reasoning is all well and good, but you might want to check your reasoning using correct variables before implying everyone else is wrong.Well I only used examples to outline my reasoning. My original conclusion is not affected by the validy of my arguments (since it were examples only). My conclusion still holds.

[Chryx]

Please!! Even without having read this I could deduce this. If you have only 64 bit adresses an array of 32 bit Ptr's does not work. Take a look at: http://arstechnica.com/cpu/03q1/x86-64/x86-64-4.html You can see that the Athlon 64 even has a whole logic-area full of transistors to support Legacy Mode. It all comes down to logic and a good memory (march 2003)

You've got it the wrong way around still, K8 is a K7 with 64bit extensions strapped on, it doesn't have 'extra logic' for 32bit mode, it has extra transistors for 64bit mode, it wasn't designed from the ground up, it was adapted from an existing architecture.

It's still an irrelevant example anyway, because PowerPC has ALWAYS been a 64bit arch, even if the consumer implementations of it haven't been, whereas the entire architecture gets an overhaul in the move from 32bit x86 to 64bit x86, the only notable difference between 32bit implementations of PPC and 64bit is the native integer word length and memory address space.

OK, the PS(1/2) is MIPS. I got confused with RISC, which they both are.

RISC/CISC is a fairly meaningless seperator these days really, almost all modern micro-architectures break instructions down into micro-ops

Well I only used examples to outline my reasoning. My original conclusion is not affected by the validy of my arguments (since it were examples only). My conclusion still holds.

That simply doesn't hold, if the baseline material is wrong, no conclusions brought about from that can be accurate.


That, and a PowerPC + VMX cpu that can't actually run PowerPC or VMX code... ISN'T a PowerPC + VMX cpu.. inherently, if it is incompatible with those instruction sets, then it isn't a CPU implementing that ISA.

QED

[aussie_geek]

I don't think the Cell is quite so limited. I've read through this fantastic explanation:

http://www.blachford.info/computer/Cells/Cell1.html

It's the future of CPUs, Apple would be foolish not to work with IBM to modify it to make the fastest freakin' Mac ever (or PC for that matter.)

Screw the talks with Intel, they've already got an "in" to the next best thing.

Joe


Excellent read!! Looks like there are great things coming around the corner.

aussie_geek

[isgoed]

You've got it the wrong way around still, K8 is a K7 with 64bit extensions strapped on, it doesn't have 'extra logic' for 32bit mode, it has extra transistors for 64bit mode, it wasn't designed from the ground up, it was adapted from an existing architecture. I have provided a hard evidence link, by deduced reasoning emphasised this evidence and still you don't get it? I even wanted to point out to you the K7 and K8 cores, but since you do it yourself I am absolutely baffled that you keep persisting that the K8 is a derivative of the K7. THE K8 IS A FREAKING NEW DESIGN! You better admit I am right. That would be less an embaressment than how you portrait yourself now. I at least know when I am wrong (The playstation example).
It's still an irrelevant example anyway, because PowerPC has ALWAYS been a 64bit arch, even if the consumer implementations of it haven't been, whereas the entire architecture gets an overhaul in the move from 32bit x86 to 64bit x86, the only notable difference between 32bit implementations of PPC and 64bit is the native integer word length and memory address space. LOGIC: the purpose of an example is not its relevance but its analogy.

That simply doesn't hold, if the baseline material is wrong, no conclusions brought about from that can be accurate. Of which your own sentence is a wonderfull example. You incorrectly identify my baseline material, so your conclusion is wrong. My baseline material is not the EXAMPLE of 64bit-x86 (which is even correct), but my baseline material lies at the heart of logic and is:

If an entity is a derivative of another entity does not mean that those entities provide the same behaviour.

Or in this context:

That Cell is based on PPC architecture does not mean that it can run PPC code.

That, and a PowerPC + VMX cpu that can't actually run PowerPC or VMX code... ISN'T a PowerPC + VMX cpu.. inherently, if it is incompatible with those instruction sets, then it isn't a CPU implementing that ISA.This is what I am trying to say all along. So I take the credits for this. Thank you very much.QEDWhat have you exactly demonstrated? It sure aint your right.

[Chryx]

I have provided a hard evidence link, by deduced reasoning emphasised this evidence and still you don't get it? I even wanted to point out to you the K7 and K8 cores, but since you do it yourself I am absolutely baffled that you keep persisting that the K8 is a derivative of the K7. THE K8 IS A FREAKING NEW DESIGN! You better admit I am right. That would be less an embaressment than how you portrait yourself now. I at least know when I am wrong (The playstation example).


Except you're wrong now, the K8 design is based off of the K7 design, heavily modified, sure, but not an entirely new design.

To use another example:

PowerPC 750 = 2+branch issue <
PowerPC 7400 = 2+branch issue, lower latency FPU, Altivec <
PowerPC 7450 = 3+branch issue, much longer longer pipelines

K7 = 32bit, 3 issue, 128KB L1 cache, 9 execution units (3/3/3 decoder/integer/FP split)

K8 = 32/64bit, 3 issue, 128KB L1 cache, 9 execution units (3/3/3 decoder/integer/fp split) - slightly longer pipeline

If you run CPU bound (fits in cache) 32bit code on a K8 and a K7, clock normalised they'll typically perform IDENTICALLY.

I'm not about to admit you're right, simply because, well, you aren't. K8 is a heavily modified K7 design, not a design from the ground up.

[isgoed]

Except you're wrong now, the K8 design is based off of the K7 design, heavily modified, sure, but not an entirely new design.

To use another example:

PowerPC 750 = 2+branch issue <
PowerPC 7400 = 2+branch issue, lower latency FPU, Altivec <
PowerPC 7450 = 3+branch issue, much longer longer pipelines

K7 = 32bit, 3 issue, 128KB L1 cache, 9 execution units (3/3/3 decoder/integer/FP split)

K8 = 32/64bit, 3 issue, 128KB L1 cache, 9 execution units (3/3/3 decoder/integer/fp split) - slightly longer pipeline

If you run CPU bound (fits in cache) 32bit code on a K8 and a K7, clock normalised they'll typically perform IDENTICALLY.

I'm not about to admit you're right, simply because, well, you aren't. K8 is a heavily modified K7 design, not a design from the ground up.Keep your pride, I don't care.

[Chryx]

Keep your pride, I don't care.

I'll take that as acknowledgement that you know you've been shooting your mouth off from a position of ignorance.

p.s. IBM have Cell related stuff under the Power architecture section on their website, since PowerPC/AS is the actual ISA in modern POWER chips.. that pretty much seals the deal, you've been wrong for the entire thread. good day.

[blitzkrieg79]

Make love not war please... Anyway that reminds me, has anyone heard any rumors on the development of IBM Power 6? I wonder how much the Power 6 will borrow from the CELL design... I mean if CELL really is what it is cracked up to be then why would IBM develop a nonCell or nonCell related processor, I know Power series are supposed to be server processors but then why is IBM pushing this CELL everywhere thing and opening up the CELL architecture to suit individual needs...

I guess there is nothing worng with having two totally different processor lines but then again IBM recently had in plans to lay off a couple of thousands of people because they had to "restructure" so I guess they are looking everywhere to save as much money as possible and to borrow the CELLlike architecture in developing the Power6 certainly would seem to save them some money. Also I have read somewhere that Power6 is supposed to sport "large frequency enhancements" and CELL (or CELL derived design) definately suits that...

[bokdol]

wow that was the most amazing viewing of geek war, i have ever seen. will not really, but it was a fun little banter.

[Chryx]

wow that was the most amazing viewing of geek war, i have ever seen. will not really, but it was a fun little banter.

Learn to use the geek side of the force, and you will achieve a power greater than any wannabe.

:p