More Cell Processor Details And First Pictures

slashflood writes "After reading two articles on slashdot about the Cell architecture and another one that criticizes the extensive roundup of the STI patents, I found the first pictures of the Cell core. It seems that at least some predictions were true. Seeing is believing." mtgarden points to this ZDNet article which says that the "first version of the chip will run at speeds faster than 4GHz. Engineers were vague on how much faster, but reports from design partners say 4.6GHz is likely. By comparison, the fastest current Pentium PC processor tops out at 3.8GHz." (More below.)

Hack Jandy writes "Anand Shimpi has some details about the upcoming Cell processor (PS3) in his personal blog. According to Anand, "Rambus announced that the new Cell processor uses both Rambus XDR memory and their FlexIO processor bus. Because Rambus designed the interface for both the memory controller(s) and the processor interface, the vast majority of signaling pins are using Rambus interfaces - a total of 90% according to Rambus." Hasn't Rambus been showing up a lot again recently? The fact that Cell uses XDR has been widely speculated, but the fact that it will also use the Rambus bus signalling is something completely new."

Comments

Pictures?

[vurg]

How about HL2 benchmarks?

Re:Pictures?

[conna01]

Because us gamers need the fastest setup possible. I hope they put these chips in everything.

Cell

[ryanmfw]

Cell processors could really dominate. With how cheap they arespeculated to be, their distributed processing, and their all around speed, the could take over a significant part of the computer marketshare. If Cell processors also have the Power4 processors in them, this could be a replacement for x86. Could be. As other articles have pointed out, x86 has had superior competition in the past, and has been able to weather it. We shall wait and see. Cheers

Re:Cell

[ryanmfw]

Er, yes, it does have a Power processor. So, it certainly could replace x86, and also usher in a new day of Linux! The year of desktop Linux has arrived! j/k, but I doubt Apple will be selling Mac OS X Cell Edition anytime soon, so, what would desktops run?

Re:Cell

[hattig]

From http://www.aceshardware.com/forums/read_post.jsp?i d=115121622&forumid=1

CELL is a Multi-Core Architecture
Contains 8 SPUs each containing a 128 entry 128-bit register file and 256KB Local Store
Contains 64-bit Power ArchitectureTM with VMX that is a dual thread SMT design - views system memory as a 10-way coherent threaded machine
2.5MB of on Chip memory (512KB L2 and 8 * 256KB)
234 million transistors
Prototype die size of 221mm2
Fabricated with 90nanometer (nm) SOI process technology

We're talking about a single-core POWER5 design (because of the SMT).

But 221mm^2 ... that's big, bigger than a 130nm Opteron, bigger than a dual-core 90nm Opteron. But wait for 65nm, and you've got something of a manageable size to make a cheaper console. I don't see 4 Cells in a PS3 though, not even at 65nm, unless it is going to cost a boatload. Still, Sony aren't a little company, I'm sure they could sort it out.

Still, I guess this means the next PowerMac G5 will be using processors with SMT finally.

Re:Cell

[ryanmfw]

Meh, I'm sure Sony would make up the $1000 defecit on every console through video game sales... And a POWER5! I am seriously impressed. Let me tell you, if the PS3 does not come with four Cells, I will be buying a few extra! I hope it has an expansion slot kind of thing, which, with the distributed architecture, is fairly feasible.

Re:Cell

[bonch]

The Cell chip uses the Power architecture, so one wonders if Apple isn't going to ship a 3.0 Ghz G5 after all and just wait for Cells instead (4.3 Ghz dual Power Mac Cell? Geez...).

Re:Cell

[radixvir]

With how cheap they arespeculated to be

Please explain why you think this will be cheap. Everything i see points to a very expensive chip. With rambus memory technology, an ibm design, and the fact that it's brand new, I dont know where you are coming up with the idea this thing will be cheap. Not to mention everyone thought the Itanium would spell death for x86, but that went nowhere.

Re:Cell

[Anonymous Coward]

There is really no reason for Linux to use a 4.6 GHz processor though.

You bite your fucking tongue!

Re:Cell

[WindBourne]

While there is no reason to run Linux on a fast CPU, ppl like to see response. In addition, so do all software developers. Must *nix developers get it from a fast kernel. Windows gets it from bigger and bigger hardware. But customers equate bigger hardware to faster response time; witness mac vs. windows and how often somebody from windows looking at mac will complaign about the slow processor speed, but does not realize that the app is actually faster on it.

Re:Cell

[doormat]

234M transistors @ 90nm is actually about as big as most graphics processors are. They tend to be 150M-200M @ 110nm or 130nm. I dont see it being terribly difficult to fab really,

Re:Cell

[ryanmfw]

Well, it seems like it's being produced on a .3 meter process! "The firms expect to begin production of Cell processors at IBM's 300mm plant in New York and Sony's Nagasaki fab later this year." Why do I sense that the article is wrong? Maybe there is a grain of truth though, and I had heard about a 30 nm plant, but I could just be crazy. :)

Re:Cell

[Wesley Felter]

The feature sizes are 90 nm and the wafers are 300 mm in diameter.

Re:Cell

[Slashdot Insider]

But 221mm^2 ... that's big, bigger than a 130nm Opteron, bigger than a dual-core 90nm Opteron. But wait for 65nm, and you've got something of a manageable size to make a cheaper console.

I don't expect the big core to make Sony back off. The original PS2 core was over 300mm^2 on (IIRC) a 250nm process: a big factor in the endless shortages of the console at the beginning of its life. It wasn't until Sony/Toshiba shrunk the process to 180nm did the shortages begin to ease.

Re:Cell

[ryanmfw]

I think you misunderstood my point. Pray tell, when do you think Microsoft will port Longhorn to the Cell? It's a product of Sony afterall, it's big rival in the consumer home devices market. Will Apple? Possibly, but not in the near future. What's left? The *NIXes!

Re:Cell

[spitefulcrow]

With the amount of power that you could get out of a few Cells you could very easily emulate a very fast IA32 CPU.

Re:Cell

[sleepingsquirrel]

300mm is the size of the silicon wafer, not the size of anything on the chip.

Re:Cell

[ryanmfw]

Ooops. Thank you for the correction. I am in serious doubt anyway of what I wrote. I had heard about 35 nm processes though, but I doubt the Cell would be manufactured on that yet. Look to the future! ;)

Re:Cell

[mboverload]

They wont, I'm just thinking aloud.

Re:Cell

[ryanmfw]

Yes, tell me, how is Sony going to sell a $3000 gaming console to more than 50,000 people? Considering that it costs tens of millions, if not hundreds of millions, to design a new chip, they are not going for the small high end market. Chances are, they're either taking a loss, or have found a way to make them for less, but considering the power, even $1000 would be cheap. The Itanium? That *did* cost $3000 and was meant to replace x86 in the server! Then, it turned out to have horrible performance. If Cell processors have horrible performance, I will be disappointed, and I would certainly retract my statements.

Re:Cell

[nokiator]

As usual, media is making a bigger deal out of this than what it really is worth. After looking at the details in the articles linked above, Cell looks to me like a combination of two well known technologies: SIMD units integrated with a microprocessor and MIMD geometry engines that is used in all modern GPUs. STI team must have figured out that moving the geometry engines from the graphics coprocessor to the main CPU may provide performance benefits in terms of processing 3D data structures. As for the 4+ GHz clock speed, this is more likely the pipeline clock speed for the SPEs and the embedded PPC core on this device would probably run from a much slower (2.4GHz?) clock. Current Intel P4 processors use 1.5X core clock in some parts of the floating point unit, so you should consider a 3.8GHz P4 to be a "5.7GHz" chip to make a fair comparison to the speculated clock speeds for the cell chip.

Re:Cell

[Anonymous Coward]

There isn't much info on this processor yet, but from what I've heard about it, I conjecture that its design is in danger of violating Nakamura's law of quantum molecular finitism, especially as the clock speeds are increased. This could result in an asymmetric shift of the lattice substrate, in which case the transistor conductivity would actually start to skew in the direction of anticonductivity (the inverse of superconductivity), forming insulating barriers. As insulating barriers would form and more heat would be generated, unbounded oscillations the molecular level could cause regenerative superhetrodyning - a cascading effect leading to the processor eventually failing catastrophically while emitting a sound remarkably similar to the Love Boat theme. Or not.

Re:Cell

[Anonymous Coward]

Depends on how much it costs Sony to manufacture these chips on their own 300mm wafer production lines.

The 90nm size is quite large, maybe it would cost around $100 a chip to make, making a PS3 around $600 to make. However that is if Sony were to release the PS3 this year. Because it is all in-house, it'll probably cost less per chip to make a Cell. When it hits 65nm, it'll cost even less because the chip is half the size (which actually means less than half the cost to make because yield increases). Then you don't have packaging costs, etc, so that is even less. Maybe Sony will get the cost per Cell down to $50 by the time of release for the PS3, meaning a manufacturing cost of $400.

Absorbing a $100 per console isn't too bad, so it could probably be sold for $299 by late 2006.

Re:Cell

[Thagg]

We think it will be cheap because

1) There are supposed to be 4 of these chips in a PS3

2) The PS3 has to cost less than $500 for it to be at all competitive

So, even if the rest of the box was free, the chips can't cost more than $125.

Thad

Re:Cell

[Flower]

They'll be cheap because Sony will take a loss on the console and make the money off of the games. Just like MS did with the Xbox. The question is just how much of a loss on the unit can Sony accept?

Re:Cell

[IdleTime]

There is really no reason for Linux to use a 4.6 GHz processor though.

This is the biggest rubbish I have ever heard!

I see a gazillion areas that requires speed and processing power. Just because you don't have the need does not mean others don't.

How someone could moderate that rubbish as insightsful is a mystery!

Re:Cell

[Fulcrum of Evil]

They'll be cheap because Sony will take a loss on the console and make the money off of the games. Just like MS did with the Xbox. The question is just how much of a loss on the unit can Sony accept?

Why whould they do that when they haven't had to before. It's not like they're anything but the preeminent console maker of the last several years. You think Sony would get away with loss-leader consoles when we have all those anti-dumping statutes?

Re:Cell

[Wordsmith]

Yes it can. At least at first, consoles are made and sold at a lost - that is to say it costs more to make and distribute one than the console company makes back on the purchase. The console company eventually makes its money in game licensing fees.

Eventually, the costs of manufacturing come down (as they always do) and the console company might actually make a profit on the systems themselves. But that takes a while (and they usually bring the selling price down a bit too, so maybe they don't make the cash back there).

If console companies didn't use that sort of a model, the PS2 would have easily been a $1,000 or more machine at debut.

Re:Cell

[Glonoinha]

Sure, it doesn't serve up a site or anything

Amazing how fast those i386 processors were at doing absolutely nothing at all.
Assign your Linux box a task or two and all of a sudden faster CPU's become appealing.

My C=64 was a bad motherfucker, right up until the point I wanted to do some serious number crunching on it (or play games.) The minute I decide that there's more to life than interacting with the operating system on an 80x25 character wide CUI ... corporate grade relational databases serving up data to a few dozen concurrent users for example, or multivariant calculus and diff/equations - and I'm looking for all the horsepower I can get.

Plus I bet it plays a mean game of Doom III.

Re:Cell

[DeadScreenSky]

Sony is taking a huge loss on the PSP (set to be released in the US in March, out in Japan since late November or so). I personally see no reason why they won't do the same for the PS3 (especially if they end up needing to include a hard disk).

(The PS2 was also sold at a very slight initial loss in the US.)

Re:Cell

[oscast]

"but I doubt Apple will be selling Mac OS X Cell Edition anytime soon"

Why not? The chip is totally PowerPC native.

Re:Cell

[ryanmfw]

Er, was that sarcasm? First, it's a POWER5, one above a POWER4, which the G5 is based on, which might make the instruction set compatible, does not mean it can make full use of the other stuff, like SMT, nor, the special stuff like the vector processors.

Re:Cell

[Fulcrum of Evil]

Any links to back that up? The only confirmed unit loss consoles I know of are the Xbox and the dreamcast. Everything else (to my knowledge) has been profitable. I looked around and the only sites that are claiming that the PSP is sold at a loss are 1up.com and some people on the chat forums. I also found a bunch of posts claiming that Sony expects to not sell the hardware at a loss. Perhaps the initial allocation was sold cheap to create buzz?

Re:Cell

[ryanmfw]

Correct, as do P4s. I believe there is a slight bit of difference between a G5 and a Cell. Just a *tad*.

Re:Cell

[ryanmfw]

I think that's what I heard actually. Damn you people for continually correcting me! Er, really, thanks!

Re:Cell

[Ohreally_factor]

How someone could moderate that rubbish as insightsful is a mystery!

I deduce that we must be on Slashdot, my dear Watson.

-- Sherlock Holmes

Re:Cell

[Donny Smith]

How the fuck is that interesting? It's outright stupid.

>With how cheap they are speculated to be
Have you seen or heard the price?

>If Cell processors also have the Power4 processors in them, this could be a replacement for x86.

What!? In what what is PowerPC a replacement for x86? If anything, it's the other way around.

x86 doesn't have to compete with PPC (and with this amoeba either) for the simple reason that Microsoft isn't likely to port Windows to it for another five or so years (at least).

Which leaves you with the still tiny UNIX/Linux market - say the Amoeba takes a 100% market share in this space, which will make it probably #3, just behind AMD.

All in all what you say here is all wrong.

Re:Cell

[Will Fisher]

I'm skeptical.

My Geforce 6800GT has 220M transistors on a 130nm process, and manages no more than around 400MHz.

How on earth can a 240M transistor chip on a 90nm process achieve 4600MHz?

NOBODY has seen even a 2 fold clock speed increase in the move from 130nm to 90nm let alone 11 fold. I know that these two chips arn't exactly comparable, but 4.6GHz?

This sounds like marketing hype to me.

Re:Cell

[Alan Partridge]

Oh come on, do you think IBM just threw away all of their prior designs when they mapped out the Cell? If those SPEs aren't largely based on the existing PowerPc family vector unit designs I'll eat my hat.

Sure it'll be different to Power 3, 4 and 5. More different than the MPC 74XX is to the PPC 601? Maybe not.

Re:Cell

[Breakfast Pants]

Your geforce does massive parrellel computations. At 400Mhz it is getting a lot done. You don't quite understand Mhz. Current Athlons are way under the clock rate of P4's yet offer similar performance. A P4's FP unit runs at more than 4Ghz. Modern DSP's run above 10Ghz. Clock rate is not the only factor in a chip. You say you know they aren't exactly comparable... you are completely right, except replace "exactly" with "even a little bit".

Re:Cell

[Will Fisher]

I thought that the idea of the cell was that it was massively parrellel. If it not, then thats many millions of transistors wasted.

Athlons do offer better performance than P4s but thats entirly down to better pipeline design.

The idea of having, what is it, 9 FPUs on a chip all running IN PARRELLEL at 4.6GHz still seems like it will have truly massive power requirements.

Re:Cell

[mike260]

They'll be cheap because Sony will take a loss on the console and make the money off of the games.

This makes even more sense when you consider the network effects - getting cells into as many households as possible will presumably encourage takeup of other cell devices.

I'm not sure I fully understand this distributed processing thing yet though. Could someone elighten me as to
- how a game console could distribute extremely time-sensitive tasks over a LAN
- how an HDTV would benefit from additional processing power
- why distributed computing needs hardware-level support in the first place

What I'm getting at is, consumer devices need to work properly as shipped, they can't rely on your toaster to provide additional horsepower. I don't much like the idea of Quake 7 dropping to 10fps whenever a housemate boils the cell-enabled kettle for a nice cup of tea.

Sharing CPU between household devices seems to make a lot less sense than selling surplus household CPU cycles to large customers for batch-type tasks.

Re:Cell

[mike260]

Actually, until recently Sony had always sold consoles at break-even price; even at launch it wasn't making a loss on the PS2. PSP is a break with this tradition but it remains to be seen which model PS3 follows.

Re:Cell

[bradleyland]

I agree wholeheartedly. Not to mention, we've now got Slashdot perpetuating the GHz myth.

"...reports from design partners say 4.6GHz is likely. By comparison, the fastest current Pentium PC processor tops out at 3.8GHz."

Re:Cell

[master_p]

But what about the compilers? will it be easy to map languages like C in a parallelizing environment like the Cell? would it give rise to functional programming languages that can be easily parallelized? will programmers adapt?

Re:Cell

[EvilTwinSkippy]

The WILL be putting out at least an Xserve with the 3 Ghz G5. The Cell is nice, but it's not designed to efficiently handle large multi-user environments.

It may be great for video editing, or 3d graphics processing, but the G5 is a lot better at walking and chewing gum at the same time. A large multi-user database, or a massive file server won't use the 8 64bit FPU's. But it will sure as hell use the dual instruction units.

(Pats shiny new G5 xserve)

I'm not saying the G5 or the Cell are bad. They are just designed to do different things. You wouldn't use a big rig truck for a Taxi, and you wouldn't use a Sedan to carry freight.

Re:Cell

[Lars T.]

x86 doesn't have to compete with PPC (and with this amoeba either) for the simple reason that Microsoft isn't likely to port Windows to it for another five or so years (at least).

XBox 2.

Re:Cell

[SunFan]

Steve sells PowerMac Cells by the seashore.

Re:Cell

[bonch]

The current rumors are that a Cell chip would be used alongside a normal G5 to aid in processing.

Re:Cell

[bbrack]

234M transistors is not the problem. As the grandparent pointed out, the primary problem is the die size - a secondary problem could be the odd floorplan (long and narrow), as this could cause greater cross die process variation Depending on defect density, 221mm^2 die could limit cell to somewhere near 30% yields - even with 300mm wafers, you'll still only be yielding 60-80 functional die per wafer - by comparison, a Pentium/Opteron, which is ~1/2 the size could yield 300-400 die per wafer. This chip will most likely be fairly difficult to fab - at the least, manufacturing costs will probably be at least 4x what they would be for other mainstream processors

Unsure

[networkBoy]

Is the cell processor a vector based CPU like previous PS cpu's? If so a cluster could prove quite formidable indeed!
(Seriously, vector processors are great for weather and nuklear simulations)
-nB

Re:Unsure

[Anonymous Coward]

"Nucular". It's pronounced "nucular".

PS3

[clean_stoner]

The Cell is going to be in the PS3, so does that mean that the PS3 will be clocked at 4.6 GHz+? That seems like a big leap considering consoles are normally running a little slower than "good" computers at the time they come out.

Re:PS3

[Dormann]

It's less of a leap when you consider the speculated release date of PS3.

Re:PS3

[satoshi1]

Late 2006?

The Xbox is 750ish megahertz. When it came out, I was gaming along on my 2ghz computer. That's quite a bit of a difference. However, with the way processors have been lately (little to no increase in speed), the PS3 will probably close that gap between console and PC speeds quite a bit when it finally comes out. I doubt processor speeds will increase enough to create the gap that exised between the Xbox (numerically the fastest console) and decent computers of the time.

Re:PS3

[Dormann]

2GHz when the Xbox was released? I'm impressed. The machine I was using to develop Xbox games at the time wasn't at 2.

I guess it's all going to boil down to whether your prediction is accurate about processor speeds stagnating. We have a solid 18 months or more to wait before PS3 is unveiled.

These are also early numbers. If in the coming year Sony sees that their system is going to be unusually ahead in processing power, they will likely loosen (cheapen) their manufacturing standards and clock down the chip.

Why does "18 months" sound familiar?...

Re:PS3

[-kabammi-]

Yes, by today's standards 4.6Ghz seems fast. Recall though, how advanced the PS2's graphics seemed when it was first announced - it exceeded PC graphics at the same time.. same with the nvidia Xbox chipset when announced was far more powerful than PC graphics at the time. In 2007 when this unit is released, the actual performance of the machine will probably be quite mainstream. This is only the announcement of the CPU guys, remember. Also remember that the PS3, if anything like the PS1 or PS2, may have a lifespan of 5-10 years. So you have to imagine that in 2017 how fast a 4.6ghz clock speed will be at that time. Who can truly imagine how fast PC's will be by then!!?! :) These consoles have to be armed with the best possible components in order to give them longevity and flexibility in the market to compete with Xbcube units produced by Nintendosoft in 2010. ;)

Re:PS3

[AssFace]

The GHz figures mean nothing at all in terms of performance expectations unless you are comparing within the same processor family.

4GHz cell != 4GHz P4 != 4GHz Opteron != 4GHz G5

Re:PS3

[Jessta]

Yay! 3
Someone who isn't being fooled by intel's crazy marketing.

Re:PS3

[Walkiry]

The GHz figures do mean something when you consider that the P4 is having trouble hitting the 4 GHz. Somehow they've managed to get a processor that runs stable and cool enough at those speeds, and that's something to be reckoned.

Umm... this was posted under Games?

[itistoday]

I understand the chip will be used in Playstation 3, but it will also likely be used in future Apple computers, of which, the G5 is already based on the Power architecture. Maybe IT would've been a better section to put this under?

Re:Umm... this was posted under Games?

[bonch]

Gamecubes also use PowerPC chips, but Nintendo is being so tight-lipped about the next-gen "Revolution" system that nobody knows what its specs will be.

Re:Umm... this was posted under Games?

[javaxman]

it will also likely be used in future Apple computers, of which, the G5 is already based on the Power architecture.

That, kind sir, is wild speculation at this point in time. Not entirely unfounded wild speculation, mind you, but we _know_ it's going to be used in the PS3. We're _guessing_ that Apple _might_ be able to use it.

I promise there'll be a PS3 before there will be a Cell-based Macintosh. So, Games it is, since, oddly enough, there's no Hardware section...

Re:Umm... this was posted under Games?

[steeviant]

Heh, you're just jealous because your computer is beige.

Re:Umm... this was posted under Games?

[Fulcrum of Evil]

Nintendo is being so tight-lipped about the next-gen "Revolution" system that nobody knows what its specs will be.

How is this a surprise? nintdo never talks about what it's doing - it's like a cult or something...

We flame Intel for touting speed...

[X43B]

I'm waiting to see how much work it can actually do before making a judgement. At the least it always exciting to have another option. I wonder how difficult it will be to take advantage of the new architecture.

Re:We flame Intel for touting speed...

[porksickle]

Good point.

Unless you manage to break down your systems or computational problems into eight digestible pieces and have an efficient flow of data between all the participants (especially given the tiny amounts of local RAM), you'd just have another glorified PowerPC CPU.

Speed isn't everything

[leathered]

While 4.6 GHz sounds impressive, I thought we were getting away from the notion that clock speed = performance. The Pentium 4 killed off clock speed comparisons.

I must admit the specs are impressive, but show me the benchmarks!

Re:Speed isn't everything

[MBCook]

That's true. But there are two important things here. The first is that it's at 4ghz. The P4 hasn't been able to reach that (though Intel origionally said it would happen by now). So it's all ready up there.

The second is that it's STARTING at 4ghz. It's one thing to say a chip can scale and run at some speed (again, I'm looking at you Intel), but to debut it running faster than the fastest mass produced CPU in the world is something all together different.

Cell should be quite formidable, and I think it will be quite interesting to see what comes of it. I've held the opinion for a few years that computers would move to having a couple of CPUs each running their own task (like in Cell), with one main (quite possibly slower) CPU controlling them all and running the OS (traffic cop, again like in the Cell). While the individual processing units are not general purpose (they are more vector oriented), it should still be interesting to see what comes of this. After all, most things people use high-end CPUs for are (or can be) vector ops, right? Compression, 3D, etc. Wordprocessing and spreadsheets don't tend to need much power. A large generalization, I know, but still... the introduction of the Cell (especiall the way it should be able to "group" its self with other Cell processors in your house) should prove quite interesting even if it turned out to be a failure (which I SERIOUSLY doubt.)

Re:Speed isn't everything

[TexVex]

The Pentium 4 killed off clock speed comparisons.

No, that was the Athlon.

Re:Speed isn't everything

[ryanmfw]

We're talking about a processor(POWER5, a step above the POWER4, which the G5 is based on) here though that is generally low in the clock speed wars, and yet performs as well or even better than higher clocked competitors. Now it's faster! It also has the huge number of vector processors, with 128, 128 bit registers.

Re:Speed isn't everything

[jc42]

I thought we were getting away from the notion that clock speed = performance.

Actually, there's a good use for such comparisons: It tells you that the writer is clueless.

I'd already read enough about the Cell to know that it's more like the PowerPC than it is like an Intel cpu. So, when I read the comparison of its supposed speed and a Pentium's, I immediately knew that the writer hadn't a clue.

Any info around about benchmarks? Those can be misleading, too, in the hands of the wrong marketer. But with enough of them, it's a lot more likely that you can glean some actual speed info.

Re:Speed isn't everything

[Prehensile Interacti]

That is why the Cell is a multi-core system.

It is not just one processor clocked at 4GHz, but is 9 of them. With the power core capable of running dual-threaded, making a total of 10 threads that can run on a single chip.

In short - that's 40GHz to you sir

Re:Speed isn't everything

[Xugumad]

In particular, for a processor whose architecture isn't even particularly similar to anything else out there, it's even less meaningful than it might be otherwise!

Re:Speed isn't everything

[Tyler Eaves]

No it's not. It's 4ghz x 10. Most tasks are NOT parrelell in nature, and saying it's equivilant to a single 40ghz chip is incredibly misleading.

Re:Speed isn't everything

[chrysrobyn]

While 4.6 GHz sounds impressive, I thought we were getting away from the notion that clock speed = performance. The Pentium 4 killed off clock speed comparisons.

Nobody is claiming that clock speed always equals performance, but think about it this way -- say you have data coming in at 10 GB/s. You could either have 8 wires (and buffers and processing) running at 10 GHz, 16 wires (etc.) running at 5GHz, 32 wires at 2.5GHz (etc.), you get the idea. If the Cell architecture processes data at 4GHz, the only thing we can be pretty certain of is that the pipeline is very deep. The benchmarks you want to see will very likely be very impressive. Perhaps the speed was partially dictated by wire density and transistor sizing?

Real world, though, what does this mean? This chip is due to be a game machine. Game workloads are, for the most part, very predictable. You process an entire screen of graphics in a very similar manner every time. This means that if you get the prediction models (and compiler hints) right, your actual performance will be very high. The same supposition could be made for encryption or any other bulk vector processing, with obvious strengths according to the instruction set of the processor. General workloads, however, do not do very well with deep pipelines. They tend to prefer less of a pipeline and less of a branch mispredict penalty. Cell will be great at its intended market, but you don't use a 4+GHz chip with over half the area (guessing, looking at the pictures) tuned for vector math (the APUs have been called SIMD) to take over the PC. For that, we will still have multicore x86 and PPC chips to dream about.

Re:Speed isn't everything

[MathFox]

You forgot the most important aspect: It is a NINE WAY chip.
Nine processors on a single chip running at over 4GHz at introduction... *WOW*

Re:Speed isn't everything

[Johnno74]

As the grandparent post said, MHZ != performance.
A good analogy tell computer illiterate people is MHZ is kinda like the RPM an engine will do. Higher RPM doesn't necessarily mean higher speed.

Also, its a RISC design. it may well do LESS in each clock cycle than x86.

And aren't we close to the theoretical limit transistors can switch at? If the cell processor starts at such a high clock rate it won't have as much headroom for improvement.

Re:Speed isn't everything

[marshall_j]

You are right. Check out this article

In laboratory tests, the Cell chip reached a top "clock speed" of 4GHz, which means it can perform more than four billion calculations per second. By comparison, the fastest Intel Pentium chip is currently capable of 3.8GHz.

This difference in basic speed is not large but Richard Doherty, director of the computer industry analysts Envisioneering, in San Francisco, says Cell's modular architecture will give it a more substantial edge for many applications.

"At first blush I think it's safe to say that it will be 10 to 20 times faster than the fastest graphics cards and processors," Doherty told New Scientist. "We think it is going to revolutionise computer science for entertainment and business."

Re:Speed isn't everything

[getch();]

Here's what this 4.6GHz number tells me. First, it tells me that either the presenter at ISSCC was ridiculously vague or that the journalist is an idiot. Call me cynical, but I doubt that even the mighty Sony has been able to clock a supposed POWER4 or POWER5 core to around double its previous frequency. So it could be a new design. If this is the case, the clockspeed tells me that the core is either extremely simple or has more pipeline stages than Prescott. We've been through this before, and that wouldn't be good. For the record, I kind of doubt that scenario is accurate.

Last I heard, the 4.6GHz number was actually the tested clock frequency of the SRAM on the chip. That's impressive and all, but it hardly begins to indicate performance of the SRAM itself. For all we know, it could run at 4.6GHz but only with 20 cycles of latency.

The short version? We need more details. Try not to get caught up in the hype machine for once.

Re:Speed isn't everything

[Prehensile Interacti]

Actually this is initially designed for Sony's PlayStation 3.

Games are inherantly very parallel.

• Loads of chracters
• Loads of animations
• Loads of discreet AIs
• Loads of polygons
• Loads of lights
• Loads of sounds

The amount of interaction between individual entities in a game world is very low, and most of a game frame is spent shovelling data asafp to the screen. Once you are in your shovelling phase, then the parallelism will be completely maxed out.

There are some of the smartest programmers around making games. Those who got over the initial pain of the Ps2's 3 core architecture should find it relatively easy to scale up. What is most exciting this time is that the 8 cores appear to be identical, meaning that there will be less restrictions about what you use each for.

The o.p. was scoffing at a 4.6Ghz spec. Whilst maybe overstating the Cell a little with my reply, the escense is that he was an order of magnitude out in his understanding.

Re:Speed isn't everything

[count_zero011]

But if you think about it, if it's based on a POWER4 or POWER5 architecture, which don't currently have very high clocks, to have this new processor beat out the Pentium 4 (which was designed so that it would scale to multiple gigahertz) is even more impressive.

Re:Speed isn't everything

[drmerope]

Indeed. Even in a slow 0.18um technology, I can easily make an 8 GHz 3-inverter oscillator ring. So what?

The "chip frequency" is determined by
1) how fast can the transistors switch
2) how many FIO4 inverter equivalents (standard measure of logic complexity) there are between the latches.

#1 is just a process technology attribute

#2 is where all the magic is because it is "how much work can take place in one cycle"

#2 is commonly reduced in a technique called pipelining.

General rule: Pipelining increases throughput at the cost of latency.

Branches especially, but in other situations as well: latency becomes a limiting factor

When this happens trading against latency is a bad decision.

For any given ISA you're likely to reach this break point *somewhere*. The i386 architecture has reached it. This is because of the latency of decoding the _complex_ instructions.

A simplier instruction set => incurs less latency penalty => can be pipelined further => can achieve higher clock speeds and accrue performance benefits to additional pipelining.

Intel, though, still has probably the best process technology in the world and as a consequence if Intel were manufacturing these cell processors they'd run even faster.

But simplier instructions tend to do less work. This means you need more instructions for the same task. More instructions might code to larger memory footprints. Larger memory footprints require faster i/o to memory and larger caches to not incur performance penalties. Thus in the end you might gain nothing.

You can see this effect within amd64. Running in 64-bit mode gives you more registers, more registers should mean faster programs, but moving around all those 64-bit variables erases the benefit. (at least in compiler run-time benchmarks that I've seen).

Re:Speed isn't everything

[ArbitraryConstant]

The very fact that it scales to such rediculous clock speeds implies that it won't be doing the same general purpose tasks as other processors. I have no doubt that it will have very strong performance in the areas for which it was designed, but it is doubtful that it will be replacing current desktop chips or server chips (including the desktop chips from IBM).

In fact, the very flexible architechture of the chip combined with the fact that PCIe is becoming available means that it's likely that everyone will be able to use these as coprocessors without any other changes to their systems -- Mac or PC.

Re:Speed isn't everything

[Johnno74]

Once that was true, not any more.

Back in the days of 8086 and earlier, CISC designs like the 8086 had complex instructions that did a lot of work, and some took upwards of 30 clock cycles to execute.

When RISC archetectures came along with their "one instruction per clock cycle" approach and kicked x86 of the day over the park Intel started putting a RISC back-end on their CPUs that decoded the x86 instructions to multiple RISC-like microcode instructions, and executed more than one microcode instruction per clock cycle.

These days many most x86 instructions are executed in one clock cycle. In spite of all its problems, x86 now has the best of both worlds now (cisc and risc), which is why it has survived.

Re:Speed isn't everything

[Fulcrum of Evil]

I have no doubt that it will have very strong performance in the areas for which it was designed, but it is doubtful that it will be replacing current desktop chips or server chips (including the desktop chips from IBM).

My thoughts run exactly counter. This is the sort of thing that Sony could leverage into a general computer setup, what with the unitized compute power. With the cell architecture, a lot of complex stuff can be scaled nearly linearly, and at low cost. I'm curious to see if that pans out, especially of this shows up in a powermac. Hell, every time I see this stuff, I think of shaowrun.

Re:Speed isn't everything

[Johnno74]

glad I could help

Re:Speed isn't everything

[ciroknight]

If you ask me, they just hired some MIPS engineers to stick a bunch of "Stream", or "Continuous Vector" units onto a modern processor, to try to undercut the video card market. Think about it this way: the Xbox hyped about having nVidia graphics core, the Gamecube is sporting a nice Radeon inside, but since Sony is a japanese megacorp, they want to do everything themselves.

I'd suspect performance to be on par with a slower pentium 4 (around 2.8GHz), due to waiting on those stream processors to use the bus, and due to SMT, which automatically performance hits whatever can't use it effeciently enough. As for Playstation games, it'll probably be the fastest damned platform ever.

Re:Speed isn't everything

[jayslambast]

My money says that the 'highest' clocked part of the chip is over 4GHz, but different part of the chip actually run slower. Lots of chip now a-days (including the P4 with its double pumped alu) run at different frequencies inside the chip. I bet the powerpc core runs around 2.5GHz while the SPEs (or whatever the vector cores are called) run at 4.6GHz. If I was an IBM architect and wanted to make everyone's life easier, I'd take the power5 core, strip out a couple of things and not mess with the timing buckets instead the chip too much. Lots of reuse without redesigning everything.
Its like saying someone claiming a corolla engine runs at 10krpms, but in reality its on the smallest gear spinning at that rate.

Re:Speed isn't everything

[AbRASiON]

Bullshit, how many SMP games are there out there?

How many games which did have SMP added saw a SIGNIFICANT boost?

Re:Speed isn't everything

[ArbitraryConstant]

My point is that there's a decent chance it won't be particularly good at general purpose code, at least in its initial itteration. I think this because it's unlikely Sony will put several POWER5s in their PS3, and Cell is more or less a big vector processor added to a PowerPC processor.

If I'm right, then it certainly won't replace POWER, PowerPC, or x86 in desktop/server use. It might give Itanium a run for its money because that gets used a lot for big numerical jobs.

"With the cell architecture, a lot of complex stuff can be scaled nearly linearly, and at low cost."

Yes, but only things that are easily parallelizable. There's a lot of that on todays computers, but there's a lot that still requires old fashioned brute force.

"I'm curious to see if that pans out, especially of this shows up in a powermac."

I think there's a good chance they'll show up on other architechtures (including PowerPC) as coprocessors on the PCIe bus. But the flexible design means they can just as easily end up on PCs as coprocessors.

Re:Speed isn't everything

[zymano]

length of interconnects ?

Re:Speed isn't everything

[Pulzar]

Sony isn't doing everything by themselves. They were planning to, but once they realized that "cell" is not going to provide adequate graphics capabilities, they went with one of the two tried and true graphics chip companies -- the only one that wasn't already making chips for the competition.

Re:Speed isn't everything

[Ohreally_factor]

Faster electricity. =)

Re:Speed isn't everything

[ryanmfw]

The thing is, they are *cheap* POWER5's. That's a big advantage right there.

Re:Speed isn't everything

[__aailob1448]

Technically, it was the PII. Mhz for Mhz, Amd's K6/3dnow! series were slower than their Intel counterparts.

The field evened out with the KIIIs and then AMD took the lead with the Athlons and still keep it today. (remember, we're talking about Mhz/performance ratio, nothing else)

Of course, AMD has consistently given more power per dollar which is why geeks are partial to it.

Re:Speed isn't everything

[ArbitraryConstant]

Just because they use the POWER ISA doesn't mean they're as powerful as a POWER5 for general purpose code. An MPC7441 (G4) doesn't compare very well to a PowerPC 970FX (G5), but they're both PowerPC.

Re:Speed isn't everything

[nothings]

The first is that it's at 4ghz. The P4 hasn't been able to reach that

This is meaningless when we don't know what is at 4Ghz. If it only means certain processing components, then Intel's already hit that. Even a lowly 2Ghz P4 already has two integer units each running at 4Ghz, because the P4 has a double-pumped integer core (scroll down to 'execution units').

Re:Speed isn't everything

[advocate_one]

it's a vector processor... which means it will operate on multiple data on each instruction

Re:Speed isn't everything

[oliverthered]

True, speed isn't everything, bandwidth counts for a lot, and their putting a few of these puppies in the console.

Also when your talking about graphics or vector processing that consoles are going to perform then raw power is what matters, not how fast you can task switch or perform string operations.

Cell processing allows the console to be bad at somethings bad not to get bogged down by them..

Now, if only QT would get their act together and make their GUI code thread safe.

Re:Speed isn't everything

[rayzat]

If only it was that easy to redesign a chip. A chip more or less has to be designed to be portable, typically at the cost of performance. The time it takes to modify a chip, make timing, and add logic for a new peripheral mix is typically longer then building a new core. Sure, sometimes some units can be reused with minimal modifications, but it's a heck of a lot more complicated then a snip here and a wire there. I wish it was that simple, I might have been able to take a vacation last year.

Re:Speed isn't everything

[master_p]

And having lots of registers makes task switching even more difficult.

But here is an idea of how to overcome that: a CPU should have a special register that each bit corresponds to a general-purpose register. When a bit has value 1, it means the register has been used and therefore it should be saved in a task switch. When a bit has value 0, the register does not need to be saved. When a task switch takes place, it is only the registers that are being used are restored.

With this technique, it would be possible to have, let's say, 256 general purpose registers, and each would task save only the registers it used. Programs could offload an algorithm that uses lots of registers to a different thread.

A further optimization would be to clear the register-use special register after a task switch, so as that only the actually used registers would be saved.

Alternatively, a CPU could have instructions that would mark registers as 'being used', so as that a hardware task switcher would only save the necessary registers.

Re:Speed isn't everything

[uujjj]

I don't think so. Task switching is very rare, occurring once every 30 million cycles or so. This will have no speed benefit.

Re:Speed isn't everything

[Lars T.]

Yeah, but it also has the really Slow Integer ALU, that "handles all of the more time-consuming integer operations, like shift and rotate". Cough.

Re:Speed isn't everything

[master_p]

No, I did not mean 'task' as in 'process'. 'Task switching' as in 'thread switching', i.e. when the TSS changes (in 80x86 talk).

Re:Speed isn't everything

[ryanmfw]

And a P Pro doesn't compare as well to a P4 either. (Err, the P3 and the P4 OTOH....) :P

joint venture

[LittleGuernica]

I believe Sony and IBM and Toshiba are going to produce this thing as a joint venture, calling it "Cyberdyne" also naming the PS3 online game network Skynet, sounds promising...

Hot

[porksickle]

I don't think the final PS3 part will be clocked much higher than 3.5GHz. Otherwise it would probably involve downclocking parts of the CPU to maintain a sane thermal profile, thus making overall performance rather unpredictable. This would especially impact games, where it's all about sustainable framerates at 100% CPU utilization.

Re:Hot

[EulerX07]

I know someone already got modded down to troll for this, but for crying out loud...

Care to substantiate your apparently qualified opinion that it won't be 4+ GHz and only 3.5 GHz? Do you have any formal training or industry experience in heat transfer when related to a 90 nm chip you know nothing about. I emphasize you because the sources of articles are actually working at the companies producing the chip, which is enough to give them the benefit of the doubt.

There are three situations that doesn't qualify you as a qualified person on this subject:
1- You overclock your computer.
2- You frequent overclocking web pages.
3- You run an overclocking web page.

It had to be said. This is not informative, this is totally off the top of your head, trying to look knowledgeable about something you probably are not, but feel free to prove me wrong.

Re:Hot

[porksickle]

Whoa! You sure make a lot of unqualified, even inflammatory statements there. It had to be said.

First of all, the press release doesn't say anything about final clock speeds, just prototype results. Considering this, even Intel showed a 4GHz P4 at the IDF in 2002. That was three years ago, ancient times for all intents and purposes and look where we are today...

Sure, it's likely that a 4+GHz Cell CPU will be used sometime in the future for some applications, but I was talking about a Sony console here, a consumer device.

It will be subject to design constraints, mainly in the areas of reliability, form-factor and especially cost.

It's safe to assume that at clock speeds exceeding 3GHz the Cell chip will require some form of active cooling. The chip is subject to the same power leakage issues that all CPUs have these days, probably even more so at a 234 million transistor count. For example, at a much lower transistor count IBM's 90nm PPC 970 still doesn't run at 3GHz, let alone 4GHz. Granted, a direct comparison might not be possible, but it's a good indicator.

As a matter of fact, there are no consumer level devices available today runnning at 4GHz, not even desktop CPUs (unless you do some insane cooling).

Given all that, the question becomes, how much active cooling can Sony afford to put in an everyday living room device? The console has to be small and reasonably failsafe, too. Active cooling solutions require space and are a therefore a cost factor (see Xbox). Beyond that, if the console requires a more extensive cooling solution, the more expensive and less failure tolerant it will ultimately be.

Additionally, if we assume a $300-$400 price point and traditionally low yields for higher clocked chips at the early stages of manufacturing, I think 3 to 3.5GHz is a very reasonable, almost optimistic estimate.

Finally, that a gaming console can not afford to have any autonomous thermal management which might affect clock speeds goes without saying. It would have tremendous TRC implications. If you can't figure out why or what that means, you obviously never shipped a console game.

Rambus kills cell...

[Duncan3]

So this will just be like the last time Rambus had their hooks into a product, it will die a very expensive and slow death.

Cell may be cheap, but the RAM will be $5/MB. Sad to see IBM repeat the mistake Intel learned from.

Re:Rambus kills cell...

[ErikTheRed]

So this will just be like the last time Rambus had their hooks into a product, it will die a very expensive and slow death.

We'll see. IBM has historically been very smart (and sometimes downright ruthless) when it comes intellectual property issues. Their IP attorneys aren't referred to as "The Nazgul" for nothing...

Re:Rambus kills cell...

[jdray]

No chance that Rambus learned a lesson, then?

Re:Rambus kills cell...

[WasterDave]

So this will just be like the last time Rambus had their hooks into a product, it will die a very expensive and slow death.

You mean like the PS2 did (given that is uses RDRAM)?

Dave

Re:Rambus kills cell...

[sjelkjd]

>>So this will just be like the last time Rambus had their hooks into a product, it will die a very expensive and slow death.

Oh, you mean like the PS2? Yep, that sure died a very expensive and slow death...not!

Re:Rambus kills cell...

[Scott7477]

If IBM attorneys are the Nazgul, that must make Microsoft attorneys Balrogs...

Ideal Linux chip.

[freemacmini]

If IBM can produce these things in enough volume this could be the ideal linux platform for the future.

PS3 is expected to sell very well so the chip production might be soaked up by the game consoles but you never know.

I bet apple engineers are salivating right now too.

Promises to be interesting for intel and AMD next year.

Re:Ideal Linux chip.

[leoxx]

Why wait?

Re:Ideal Linux chip.

[ArbitraryConstant]

It's unlikely that Cell processors will have comparable performance on non-vector code to current Intel or AMD (or IBM) desktop and server chips. It's a vector processor.

Also, its flexibility means it could be used as a coprocessor on existing systems -- Mac or PC.

Re:Ideal Linux chip.

[be-fan]

If the POWER chip in the thing runs at a decent clockspeed, it should.

Re:Ideal Linux chip.

[ArbitraryConstant]

I believe it does use the POWER ISA, but that doesn't mean it'll be fast. For example, I doubt Sony is going to be using a full POWER5 with SMT in their playstations.

More likely, the chip will be binary compatible with POWER but lacking in performance for general purpose code.

More likely, servers and desktops will continue to use AMD, Intel, or IBM processors with the option to use Cell coprocessors in PCIe cards for applications that need them.

Re:Ideal Linux chip.

[rayzat]

Chip supplies might not necessarily be soaked up by the PS3. Every wafer produces an assortment of chip speeds, and since the goal for every game box is to be identical, overly fast or slow chips might get out. This chip probably isn't going to be the ni plus ultra of desktop linux computing, but it could potentially be very useful in an academic setting. Where you can have some massive vector processing for simulations( biology, chemistry, etc. ) and a CPU for running a useable OS.

Some specs from Sony press material

[Anonymous Coward]

http://www.scee.presscentre.com/imagelibrary/detai l.asp?MediaDetailsID=25555
:

CELL...bringing supercomputer power to everyday life with latest technology optimized for compute-intensive and broadband rich media applications

SUMMARY:

Cell is a breakthrough architectural design -- featuring 8 Synergistic Processing Units (SPU) with Power-based core, with top clock speeds exceeding 4 GHz (as measured during initial laboratory testing).

Cell is OS neutral - supporting multiple operating systems simultaneously

Cell is a multicore chip comprising 8 SPUs and a 64-bit Power processor core capable of massive floating point processing

Special circuit techniques, rules for modularity and reuse, customized clocking structures, and unique power and thermal management concepts were applied to optimize the design

CELL is a Multi-Core Architecture

Contains 8 SPUs each containing a 128 entry 128-bit register file and 256KB Local Store

Contains 64-bit Power ArchitectureTM with VMX that is a dual thread SMT design - views system memory as a 10-way coherent threaded machine

2.5MB of on Chip memory (512KB L2 and 8 * 256KB)

234 million transistors

Prototype die size of 221mm2

Fabricated with 90nanometer (nm) SOI process technology

Cell is a modular architecture and floating point calculation capabilities can be adjusted by increasing or reducing the number of SPUs

CELL is a Broadband Architecture

Compatible with 64b Power Architecture(TM)

SPU is a RISC architecture with SIMD organization and Local Store

128+ concurrent transactions to memory per processor

High speed internal element interconnect bus performing at 96B/cycle

CELL is a Real-Time Architecture

Resource allocation (for Bandwidth Management)

Locking caches (via Replacement Management Tables)

Virtualization support with real time response characteristics across multiple operating systems running simultaneously

CELL is Security Enabled Architecture

SPUs dynamically configurable as secure processors for flexible security programming

CELL is a Confluence of New Technologies

Virtualization techniques to support conventional and real time applications

Autonomic power management features

Resource management for real time human interaction

Smart memory flow controllers (DMA) to sustain bandwidth

Re:Some specs from Sony press material

[jdray]

Oh, great. Future comparisons of computing capability will be based on SPUs. "Hey, I got the new 12-SPU chip in my computer. How much SPU is in your computer?"

Re:Some specs from Sony press material

[Sunspire]

Warning: Pregnant women, the elderly, and children should avoid prolonged exposure to CELL.
Caution: CELL may suddenly accelerate to dangerous speeds.
CELL contains a liquid core, which if exposed due to rupture should not be touched, inhaled, or looked at.
Do not use CELL on concrete.
Discontinue use of CELL if any of the following occurs:
* Itching
* Vertigo
* Dizziness
* Tingling in extremities
* Loss of balance or coordination
* Slurred speech
* Temporary blindness
* Profuse Sweating
or
* Heart palpitations

If CELL begins to smoke, get away immediately. Seek shelter and cover head.
CELL may stick to certain types of skin.
When not in use, CELL should be returned to its special container and kept under refrigeration.
Failure to do so relieves the makers of CELL, Sony Incorporated of any and all liability.
Ingredients of CELL include an unknown glowing substance which fell to Earth, presumably from outer space.
CELL has been shipped to our troops in Saudi Arabia and is also being dropped by our warplanes on Iraq.

Do not taunt CELL.
CELL comes with a lifetime guarantee.
CELL! Accept no substitutes!

Re:Some specs from Sony press material

[gfody]

CELL is a Real-Time Architecture
CELL is a Confluence of New Technologies

sounds like someone was playing with their execuspeak magnets

Re:Some specs from Sony press material

[Rorschach1]

CELL is a dessert topping AND a floor wax!

Re:Some specs from Sony press material

[be-fan]

The real-time architecture makes sense --- the Cell has timing circuits to enable precise control of when computation happens. It also has a real-time OS that allows programs to make demands about when computations need to be completed.

Re:Some specs from Sony press material

[jo42]

> 8 Synergistic Processing Units

There is a Dilbert comic in here somewhere...

Eight spoos, eh?

Re:Some specs from Sony press material

[Michael Hunt]

It lengthens, it strengthens... It's a friend, a companion, it's the only product you will ever need, you can drive it away today.

Re:Some specs from Sony press material

[gstoddart]

CELL is a dessert topping AND a floor wax!

Wow. I can't believe this is currently moderated as friggin' informative.

How do things like that happen?

Re:Some specs from Sony press material

[TeknoHog]

CELL is an antebellum flagellum fella.

Re:Some specs from Sony press material

[dourk]

Do not eat Cell.

Re:Some specs from Sony press material

[Tropaios]

The Happy Fun Ball is not ammused...

Re:Xbox

[demonic-halo]

I think more likely, Microsoft will just end up putting in more Cells in their next XBox if the PS3 proves successful.

I don't get it

[pHatidic]

This seems good, but what exactly can I do with a cell processor other than play games with better graphics? It seems like the vast majority of people don't use even half of the power their computers have today, and if there are bottlenecks in todays computers it is because of RAM and the OS and not because of the CPU. Other then games, when will I be able to do other than maybe look for aliens faster.

Re:I don't get it

[Locke2005]

The cell architecture is optimized for streaming things like video and audio, not for playing games. If you can't think of any applications for which high speed encoding/decoding of things like mpeg, jpeg, and mp3 would be useful, then you don't need a cell processor! This chip will see a lot more use in embedded audio/video devices than it will in desktops; and in fact it buys you nothing if all you want to do is run office software suites. But then, the latest pentium 4 also buys you nothing if all you want to do is run office software suites! In other words, any application that you would actually need the latest greatest x86 chip for, the cell will perform an order of magnitude faster (once the software is available). Note to Wintel: Be afraid... be very afraid!

Re:I don't get it

[gbulmash]

This seems good, but what exactly can I do with a cell processor other than play games with better graphics? It seems like the vast majority of people don't use even half of the power their computers have today, and if there are bottlenecks in todays computers it is because of RAM and the OS and not because of the CPU. Other then games, when will I be able to do other than maybe look for aliens faster.

"If you build it, he will come."

If you create a machine so powerful that there's nothing that fully utilizes its capacities, that merely spurs all sorts of geeks to dream about how they can push that machine to its limits, then overclock it, then put it all in a case made of Legos.

- Greg

Re:I don't get it

[beelsebob]

OK then, lets stop then, 99% of people are happy, they've got CPUs that do everything they want to do. So lets stop designing new CPUs that run faster, cooler, do more and consume less power... Lets just make do. After all, you don't see people saying "I can't get a fast enough laptop that doesn't eat the battery" and you don't see scientists wanting huge number crunching beasts, and you don't see gamers wanting faster and faster chips.

Re:I don't get it

[pHatidic]

I already know what other people will use it for, what I am asking is what will I use it for? I'm not a gamer anymore nor am I a scientist. If 99% don't have a compelling reason to by it then they won't, so it won't make money and more cool stuff like this won't be invented in the future. So my question is, what is going to make 99% of people buy it, not what will make 1% of people buy it which is what you have answered.

Re:I don't get it

[glass_window]

I thought the deal with Lenovo meant IBM could not produce anymore PCs or laptops.

Re:I don't get it

[halltk1983]

Not many people exceed 100 mph, or corner a 1g+, but they still buy sports cars... If they make it, people will buy it, if for no other reaon than to say "I have a 4Cell HP Pavilion, and you only have a 2Cell."

You seem to have forgotten that as a race, humans are egomaniacs...

Re:I don't get it

[nomel]

When I do complicated real time video processing, I need lots of power.

Re:I don't get it

[Fulcrum of Evil]

Not many people exceed 100 mph, or corner a 1g+, but they still buy sports cars...

I can do both and I've only got 140hp ;)

Re:I don't get it

[halltk1983]

That's exactly my point. You don't NEED it to do what you want... But it is a status symbol that people want. Something they can point to and say "That baby is top of the line" not realizing that something so much less could fit their nees. As long as America is trying to beat the Jones's, they will always go with the most powerful that they can.

Re:I don't get it

[rusty0101]

Fabricating and selling CPU chips does not equate to making or selling PCs or laptops.

Re:I don't get it

[Scott7477]

According to the referenced article, this chip is supposed to be able to transmit data to and from off-chip memory at 100 gigabytes per second.

The same article has this hilarious quote from an Intel Fellow: "One of the big problems with Cell, said Justin Rattner, an Intel Fellow, is that the processing units aren't identical, a situation that increases complexity and the opportunity for bugs."

This from a company whose early microprocessor designs were frequently referred to as "brain-damaged."

Re:I don't get it

[wwahammy]

RAM and OS are minor compared to the real bottleneck in general performance: the hard drive. Hard drives have access times in the millisecond while ram has it in the nanoseconds. Until someone figures out how to make a hard drive that has access times a thousand times quicker, we're still going to have processors waiting for data longer than they really should be. Research in this area is catching up but after many, many years of processor and RAM research its got a ton of catching up to do.

Re:I don't get it

[Ohreally_factor]

I must have been absent the day you were made representative of 99% of the population. Were you elected, or were you appointed by someone? Was there a party? Cinnabuns in the conference room to celebrate? Dammit, I'm always absent when they have free Cinnabuns.

Re:I don't get it

[__aailob1448]

Damn right.

Re:I don't get it

[CCRancor]

What's the first thing you'd do if you were IBM and knew you could reform the entire PC industry with a CPU for everything (games, home theater, and ordinary work) and all in a PS3-sized box?

If you're answer is "sell of the PC department" you're right.

Re:I don't get it

[samdu]

And that's what'll keep the US on top. :)

Seriously, though, the drive for the next, best, fastest thing, while having some downsides, does move humanity forward. Think of how much progress would have been lost if we didn't have that drive. Theoretically, we could all do just fine with horses, for instance. And microwaves aren't really all that necessary. Space? No real reason to go there. This is just a microcosm of that mindset that has allowed man to reach the stars.

Re:Xbox

[Thu25245]

Thing is, the next Xbox will be using a PowerPC 970. So it will share a common ancestor (POWER) with the Cell.

I wonder, how compatible are the two CPUs' instruction sets? Will Microsoft be able to drop a Cell into a future revision of the Xbox2 and maintain backward compatibility? Could someone theoretically hack a PlayStation3 to run Xbox2 games?

But...

[sH4RD]

"first version of the chip will run at speeds faster than 4GHz. Engineers were vague on how much faster" But I thought GHz didn't matter?

Re:But...

[Bruha]

They're vauge becuase they dont want to get into the ghz vs performance aruguement.

8x64bit FPU's along with a seperate processor that can handle 2 threads at once along with a full speed memory bus may be able to blow the doors off anything in production. Also considering that this memory is cheap enough to put 3 of these in a gaming ocnsole that may retail at 300 dollars points to the fact that Rambus may be able to churn out multi gigabit solid state drives to support these blazing fast speeds.

If this all pans out you'll have a set top box that's light years faster than your desktop machine with it's normal hard drives.

Of course I'll wait to see it first.

Re:But...

[ALeavitt]

No, it's MHz that don't matter. GHz... those matter. Just not as much as THz.

Re:But...

[nuclear305]

"But I thought GHz didn't matter?"

Where in that sentence did it say it mattered? It simply stated a clock speed--one of the various technical specs mentioned in TFAs

Had it said something like, "The first version of the chip will crush all competition at speeds faster than 4Ghz." then I'd agree with what you've said, but there was no relation of clockspeed to performance in any of the articles as far as I can tell.

noisy little bugger...

[Sebastopol]

My PS2 is already unnacceptibly noisy. When playing dramatic games like Silent Hill, the sublte nuances are drowned out by the WHIRRRRR of the PS2 box. I guess I should lock it up in a cabinet or something.

I can't imagine something this big and fast being quieter.

Re:noisy little bugger...

[ScrewMaster]

There are devices known as "stereo headphones" that can help with this. They've been out for a while now.

Re:noisy little bugger...

[Sebastopol]

yeah, but i can't find any headphones big enough to find around my gigantic nerd head, and my hair is so greasy they just slide off and land in my trough of cheetos...

loser.

Re:noisy little bugger...

[__int64]

"When playing dramatic games like Silent Hill"

Yeah, the disc loading sounds always blow away any surprise moments, whenever the disc starts to spin up you know something big is about to happen.
I've cycled though each of this generations systems and found gamecube to be my hands down favorite on design, hardware and especially noise factor.

How they got the die photos

[__aadkms7016]

They zoomed in on this press photo of an engineer holding a die.

Re:How they got the die photos

[TheGuano]

That chip is friggin' huge. I wonder how big the heatsink/fan is going to be.

Re:How they got the die photos

[__aailob1448]

I'm not sure she's an engineer...

Anybody remember the 25x?

[k0ft]

Chuck moore's 25x had 25 misc cores on one die. Specs:

.2 sq mm asynchronous microcomputer core 5 x 5 array of cores: 60,000 Mips 5 horizontal, 5 vertical parallel interconnect buses: 180 Ghz bandwidth Specialized computers to interface off-chip. Max power 500 mW @ 1.8 V, with 25 computers running 100mAh battery life is 1 year, with 1 computer running throttled 64-pin SOIC: mirrored pin-out to 4ns cache SRAM Array chips on 2-sided PCB

Shortly after he announced the chip, he took the link off his page. According to a post from him he's in a lawsuit

The Cell processor is pretty cool, but i see some room for improvement. They could have made simpler cores, and lots more of them.

Re:Anybody remember the 25x?

[nurb432]

I remember it... Does anyone know how the law suits are going? Would love to see this chip make it into the real world.

Re:Anybody remember the 25x?

[babazaroni]

Chuck Moore is designing a chip for a new firm called Async Array Devices. It is probably a variant of the 25x.

They are supposed to announce the chip at a microprocessor conference in April.

It appears the new chip will not be available for general purchase. They don't won't to be in the chip sales business, just provide solutions.

Check out http://ultratechnology.com/fsc04.htm.

I suspect that http://enumera.com/ is behind this new company.

The Sony hype machine strikes again

[Laconian]

Remember how the Emotion Engine worked us all into a lather five years ago? And when it came out, it was just merely competitive with contemporary processors? Sony is great at churning out nerd fetish tech, but they have a terrible track record of living up to their promises. Let's hope it's different this time.

Re:The Sony hype machine strikes again

[thogard]

Its sort of like that cool graphics chip used in the N64. That no one ever took advantage of except for the spinning N at the start of some games and I understand that was written by the SGI guys as demo.

From what I can see, all this special hardware that gets tacked on modern game systems never gets used because a game developer can either use it and lock their game to one platform or ignore it and have something they can also sell to the PC market.

Re:The Sony hype machine strikes again

[cosmo7]

A lot of the performance in games consoles comes down the the quality of the profiler and the compiler optimisations available. Sony were still improving the original PlayStation's performance after the PS2 was available.

Re:The Sony hype machine strikes again

[Queer Boy]

Remember how the Emotion Engine worked us all into a lather five years ago? And when it came out, it was just merely competitive with contemporary processors?

Yup and with a hardware bug that prevented hardware anti-aliasing.

Re:The Sony hype machine strikes again

[gabe]

You forget, it's not just Sony designing and producing the processor this time. Just look what IBM did for Apple with the G5.

Re:The Sony hype machine strikes again

[adpowers]

Just look what IBM did for Apple with the G5.

You mean promise 3 GHz for the summer of 2004 and still not deliver six months later? Oh, woops, I forgot, we aren't supposed to talk about that.

Re:The Sony hype machine strikes again

[PMuse]

Remember how the Emotion Engine worked us all into a lather five years ago?

Ginger will "sweep over the world and change lives, cities, and ways of thinking." - Steve Kemper

"The Cell processor-based workstation will totally change the digital content creation environment. Its overwhelming power will be demonstrated in every aspect of the development of all kinds of digital entertainment content, from movies, broadcast programs to next generation PlayStation games." - Sony

But wait! Now coming, the NEXT BIG THING!

Cell processor planned to be inexpensive

[Anonymous Coward]

My friend just called me from the ISSCC. He got a couple more bits of info, including that STI plan to recoup their R&D expenses largely from other consumer multimedia devices and NOT from selling Cell processors or Cell processor based computers, meaning they will be surprisingly inexpensive. Yeah!!!

Re:Cell processor planned to be inexpensive

[temojen]

or not available.

Apple will figure something out.

[demonic-halo]

Knowing Steve Jobs..

By 2006, instead of having a "Genie" effect in OSX, he'll actually have real Genies.

Re:Apple will figure something out.

[FuturePastNow]

A dual-4.6GHz G6 Powermac. I might have to sell my car for one of those (car is only worth $3000).

But flaming Intel is fun!

[Sycraft-fu]

Seriously, I've noticed that Intel is probably teh 2nd favourite target here after Microsoft (ok, 3rd favourite now that SCO claimed the throne). When Intel does something some way, it's bad and stupid, when someone else does, it's good and better than Intel.

Like the whole ISA things. People always seem to be down on Intel for sticking with x86. They like to talk about what a hack it is, and how much better a RISC archetecture is, and so on. Then the IA-64/x86-64 fight comes and now AMD's the good guys for sticking with x86-64 and pushing backwards compatibility. WTF?

But yes, you are correct, the real question is how much work can it get done, sepcificly how much work can it get done for the kinds of things consumers want. It's all well and good to crank on theoritical benchmarks, but what really matters is the wall clock. How long does it take to render a scene in a 3d program, how long does it take to create a mixdown in audio software, etc.

Only time will tell with these. However, given that IBM is casting in on them, I'd say it's more than just hype. They already have a fast processor on a "normal" technology, they aren't messing with this new stuff just for fun.

Re:But flaming Intel is fun!

[mboverload]

x86 seems to a be a pretty capable architecture to me.

Re:But flaming Intel is fun!

[Sycraft-fu]

It is, and receantly with developments in chip design and compiler design, the architecture of a chip has become much less of a big deal.

Back in the day, RISC was important because it allowed pipelining, the ability for a chip to be doing multiple things at once. Like old MIPS chips used to have 8 parallel piplines that took 8 cycles to execute an instruction, giving an effective rate of one instruction per cycle. Couldn't do that with CISC. Well now processors are decoupled from their ISAs. Each of those instructions is translated into a number of micro operations, which are actually what get handled by the processing section. Likewise it means there can be more registers than are exposed by the ISA.

The upshot is that it doesn't matter as much it used to.

However, there are still plenty of people who like to villify Intel for sticking with x86. They declare it to be an olde kludge of an architecture that needs to die and makes things all slow. However when AMD decided to stick with it, rather than hop on the EPIC bandwagon, they are suddenly heros for maintaining backwards compatibility, which is the whole reason Intel has stuck with x86 for so long.

What's I'm pointing out is the bashing is done against Intel, regardless of what they do. Intel is in the "bad" position, no matter what that is. Like with the cell chips and speed. Slasdotters have been long raging on Intel for making a design that has higher MHz but less performance per MHz (as opposed to AMD). They declare it to be a marketing gimick, etc. Now here we have an article talking about cell chips that are designed to cycle even faster, and taking shots at how slow Intel chips cycle by comparison.

It's not that these people actually have good reasons to like or dislike the decisions, they just dislike Intel and so slam on them.

Re:But flaming Intel is fun!

[steeviant]

If I were to villify Intel, it wouldn't be for sticking with the x86 instruction set.

I'd villify them for...

Building the 8086 down to such a super low cost, by crippling them with a pitifully small number of interrupts, registers and DMA channels, they knew it was a bad idea but went ahead with the design anyway.

Making a chip (the 286) with an enhanced 16-bit mode, but no way to switch back to real mode (fixed in the 386) which resulted in the 16 bit mode of the chip being nearly useless.

Failing to properly address the I/O problems of the first generation processor until the fifth generation CPU.

Forever upping x86 core speeds in an attempt to convince punters that GHz == instructions executed, despite it meaning that the chips would fail more, and make computers noisier and hotter.

And finally, copying AMDs 64bit architecture without giving any credit. :)

Re:But flaming Intel is fun!

[Detritus]

Making a chip (the 286) with an enhanced 16-bit mode, but no way to switch back to real mode (fixed in the 386) which resulted in the 16 bit mode of the chip being nearly useless.

You weren't supposed to switch back to real mode, which was obsolete. They underestimated the PC industry's fixation on backwards compatibility and overestimated the ability of Microsoft and other software vendors to produce advanced operating systems. It was a decent chip, but it wasn't what most of their customers wanted.

Re:But flaming Intel is fun!

[evilviper]

What's I'm pointing out is the bashing is done against Intel, regardless of what they do. Intel is in the "bad" position, no matter what that is.
Bullshit.

Start showing examples of THE SAME PERSON bashing Intel for sticking with x86, and cheering-on AMD for sticking with x86...

I'm willing to bet you won't find any examples, because there are two seperate sets of people expressing their own (contrary) opinions on different subjects, that you want to lump together.

Same thing goes for CPU performance/power... A lot of people want top-of-the-line performance, and a lot of people want lower-power CPUs.

Re:But flaming Intel is fun!

[master_p]

Another cool capability would be to be able to store user-defined instructions inside the CPU; these instructions would be made from simpler RISC-like instructions or from other user-defined instructions. A programmable microcode, if you will.

It would certainly make a big difference if you tell the CPU, for example, to execute a 'multiply a matrix 3x3 with another matrix 3x3' from the programmable microcode. For starters, the CPU wouldn't need to fetch the routine from RAM. The CPU would execute the command from the programmable microcode area, thus allowing for greater speed and less memory. The CPU could have a cache and prediction unit for the programmable microcode area that remembered the optimized version of the user-defined instruction, thus allowing for even greater speed.

But you are right, we are stack in 80x86 mode...

Re:But flaming Intel is fun!

[SQL Error]

Back in the day, RISC was important because it allowed pipelining, the ability for a chip to be doing multiple things at once.

The Z80 was pipelined.

Like old MIPS chips used to have 8 parallel piplines that took 8 cycles to execute an instruction, giving an effective rate of one instruction per cycle.

No they didn't. No MIPS processor has ever been built with 8 parallel pipelines.
Now, the R4000 family had one 8-stage pipeline, but that's completely different.

Couldn't do that with CISC.

Complete nonsense. CISC processors were pipelined long before designs like the Pentium Pro appeared.

Re:But flaming Intel is fun!

[b1t r0t]

You weren't supposed to switch back to real mode, which was obsolete.

Except the problem was that Intel insisted on sticking with 16-bit offsets to segment registers, which was also obsolete.

Linear addressing was the way to go, and limiting any given object in memory to 64K bytes size was one of the main problems with the 8086 through 80286 architecture. The main workaround to that was segment arithmetic, which only worked in real mode. Nobody wanted the 64K straightjacket of coding for 286 protected mode.

Re:But flaming Intel is fun!

[Detritus]

I used to run UNIX (AT&T System V) on my 80286 system. While being limited to 16-bits was a problem, too many people assumed "All the world's a VAX", it had some unexpected benefits. The liberal use of segments for code and data made the system trap many memory accessing errors that went unnoticed on systems with linear 32-bit address spaces. Unfortunately, the CPU spent much of its time reloading segment descriptors. The C compiler was poor at optimizing memory accesses.

If you look at Intel's old literature, they assumed that people would be writing software in PL/M and Pascal for a Multics-like operating system. They weren't expecting the enormous popularity of C and UNIX-like systems that used large linear address spaces.

It may be the most powerful processor yet, but....

[Sgt_Astro]

...will it still respect you in the morning?

windows..

[mottie]

with IBM is pushing linux, I can't see microsoft ditching their good buddy Intel, and throwing money at IBM by porting windows to run on the Cell processor. I would guess that this would allow Intel to retain their market share, even if the Cell is infact a far superior processor.

Re:windows..

[djvern]

...becasue its not like their ditching intel in their gaming platform is it? http://www.theinquirer.net/?article=12552

Re:windows..

[Queer Boy]

Microsoft only cares about having Windows on what people are using. If Cell workstations start becoming prolific, you'll see Windows for Cell, just like NT on PowerPC and Alpha when MS was worried RISC workstations were going to eke them out of the market.

RTFA

[temojen]

The Cell CPU has a POWER Processor with VMX (it's vector based), plus 8 stream processors (which kick ass on vector processing units for some applications). So you've got

• a regular CPU (good for program flow/logic and interdependant operations),
• a vector unit (good for large arrays with no conditionals),
• and 8 stream processors (good for applying the same operations plus flow control to lots of independant chunks of data).

w00t!

Re:by comparison...

[rco3]

what does your oscilloscope need a D/A converter for? Do you mean an A/D converter? And why 1,024 bits? That's 128 channels worth of 8-bit A/D.

What scope is this?

Re:I wonder....

[ryanmfw]

Ah, I don't prophesy it being quick. I still do see a mini revolution though. I wonder what M$ has to beat back a server processor with essentially hyper threading, running at 4.6 ghz, attached to 8 vector processors, each with a lot of registers and cache, which are using extremely fast memory, that can connect to other, similar processors nearby. I don't think they have much. :P

Re:Where the fuck are you getting this?

[itistoday]

I'm fairly sure I suggested the IT section, though I could be wrong. Perhaps lending me some of your gusto would help me see better?

Conspiracy Theory

[Sophrosyne]

fact #1 Apple and Sony have been awfully close for the past few years- with some dialogue between the two CEOs.
fact #2 Apple has signed up to display at E3 this year- but hasn't published any official info on their site.
fact #3 The Mac is somewhat deficient when it comes to gaming when compared to the Windows PC.
So my speculation is that it is possible that Apple intends to build a new Mac aimed at the gaming market that will be compatible and play Sony's PS3 games- Apple in turn could publish games for the PS3.

Re:Conspiracy Theory

[Sophrosyne]

Here is a link to the Apple E3 Article
http://www.thinksecret.com/news/0502briefly.html
Also if you remember Sony recently admitted they made a mistake with their new walkman- and you also have to take into account Japanese culture and the concept of competition are not always thought of in the same respect- especially when Apple dominates the Japanese computer market (and now mp3 market).

Re:Conspiracy Theory

[painandgreed]

,P> So my speculation is that it is possible that Apple intends to build a new Mac aimed at the gaming market that will be compatible and play Sony's PS3 games- Apple in turn could publish games for the PS3.

IMHO, this is what apple needs. If the PS3 was to act to the Mac the way XBox is to the PC, then the mac wold get the games it needs to try and gain market share over Windows. For a vast number of people, games are the killer app. I'm a huge mac fan, but if people tell me they want a comptuer to play games on I tell them to buy Windows. hardware-wise, i think they Mac could hold their own with Windows, but they need the game library to even try and become a gaming platform. With the PS/2/3 library to draw upon, they would have that.

Trouble is, I can't see how Sony would benefit. Perhaps they're willing to help Apple to hurt MS/XBox. Thus, include PCs in the console wars. After all, for every copy of Windows not sold, that cuts into the money needed to shore up the loss leader XBox. For every Mac sold that could play PS3 games, that would mean more games sold.

Re:Conspiracy Theory

[rindeee]

"the minimac of incredible cheapness"...is that would be used by "the rodents of unusual size"?

Re:Conspiracy Theory

[I judge you]

Or, Apple will be there to show off systems for potential PS3 developers since the G5 will probably be the official development system

Get a fucking clue. The offical developement system will be something pretty close to PS3 hardware - i.e. nothing like a Mac.

Re:Conspiracy Theory

[DeadScreenSky]

especially when Apple dominates the Japanese computer market

Seriously? When did this happen?

Re:Conspiracy Theory

[Queer Boy]

So my speculation is that it is possible that Apple intends to build a new Mac aimed at the gaming market that will be compatible and play Sony's PS3 games- Apple in turn could publish games for the PS3.

Or it's possible that Apple is writing the OS for the PlayStation3.

The overwhelming majority of people do not play games on computers nor do they want to. The living room is where entertainment is king. The sheer horde of developers for consoles compared to the paltry handful for PCs should be a clue as to where the money is.

Re:Conspiracy Theory

[Kenshin]

Apple is at E3 EVERY year.

An out-of-the-way room off the main showfloors, but they're there. (And usually have sandwiches.)

Re:Conspiracy Theory

[bi_boy]

Maybe then that damned Halo-RTS game would finally come out. I've been waiting years for it!

Apple's connection to the Cell processor

[bonch]

Apple's Connection to the Cell Processor

Re:Apple's connection to the Cell processor

[EvilTwinSkippy]

Actually I think that is more about Apple useing CELL in the future. Apple wouldn't license OS X to Sony.

They learned from their last attempt at clones that they make most of their money off the hardware. And if someone has a license to make knock offs of your product, they WILL try to compete with you on the high-end high-profit models over the low-end thin-margin ones.

And Sony is a high-end, high-profit brand name. You don't buy a Sony Viao because it's cheap. You buy it because it looks snazzy, and has a pile of ubilicils and software that strap to video and audio gear.

Much like a Macintosh.

Power consumption

[Anonymous Coward]

For those of you wondering about the power consumption of this thing, perhaps you should note that Sony just licensed LongRun2 from Transmeta. It is a dynamic solution for power consumption and leakage that will probably end up in the 65nm versions coming out next year. google transmeta sony for more.

Once touted as the Intel killer, perhaps Transmeta will finally have its day.

Re:Where the fuck are you getting this?

[itistoday]

Thanks for the gusto! I can now see the reason for the confusion...

"Maybe IT would've..."

I think I should've specified "IT section"....

Re:Cell Processor enhances cross-platform software

[satoshi1]

High end cames on your cellphone? Tell me when I can buy a cellphone with a Geforce 6800 PCI-e (heck, I'll settle for a Geforce 2 MX in a cellphone) in it, and then maybe we'll talk high-end games.

Cell- the first processor to run on hype

[i41Overlord]

Sounds a lot like their Emotion Engine, which could "render Toy Story like animation in real time".

Too bad it couldn't come close to living up to the hype.

With all this talk about the Cell being 4x as fast as a dual core Opteron, I'm willing to bet that when the chip comes out its actual speed will be similar to a low/mid range P4/Athlon.

It won't kill AMD or Intel, it won't dominate the processor market, it won't take the computing world by storm, and it won't even take the performance lead.

The only thing this is taking by storm is the PR and marketing business.

Re:Cell- the first processor to run on hype

[be-fan]

I'd love to see where Sony said the EE could "render Toy Story in real time", as opposed to some dittohead journalist. These are presentations at an engineering convention, not some stupid "tech" articles in USA Today. Look at the specs, evaluate the thing on its own merits. Don't make baseless conjectures (which is precisely what you're doing).

Re:Cell- the first processor to run on hype

[Anonymous Coward]

engineering convention turned to pr feeder?

yes. look at the specs. they're just that: SPECS aka numbers on paper, nothing more than that. emotion engine was also bunch of pretty theoretical numbers.

then they're even admitting that it's to the programmer to choose how the units will be used. it's going to be a huge pain.

Re:Cell- the first processor to run on hype

[gl4ss]

from couple of replies downwards.. http://archives.cnn.com/2000/TECH/computing/02/01/ emotion.engine.idg/index.html

replace ps2 with ps3 in that article to bring it up to date.

they know from experience that the empty hype works.

look, the architechture may be a very hot one, but it ain't going to be both cheap and as fast as they're flashing at us.

they could just screw gameconsoles if it were anyways.

Re:Cell- the first processor to run on hype

[be-fan]

Where does it make any statements in there that can't be backed up? The EE *is* twice as fast as a 733MHz P3 --- the only reason the XBox has better graphics is it's NVIDIA GPU. It does actually get 6 gigaflops, and it can actually draw 70M polygons per second. It does have a DVD drive, and can connect to the internet. And yeah, it's black. How is any of that hype? Certainly, none of it is at the level of what the OP was claiming, that Sony said the EE would render Toy Story in real time.

Re:Cell- the first processor to run on hype

[be-fan]

I don't get your point. Yes, IBM is citing theoretical peaks for Cell. That's standard procedure in the industry. I mean, consider all the supercomputers --- the press releases before they are built always show the theoretical peaks!

Re:Cell- the first processor to run on hype

[I judge you]

The EE *is* twice as fast as a 733MHz P3 --- the only reason the XBox has better graphics is it's NVIDIA GPU

There are so many things wrong with this, I don't know where to begin. Let me stick to the basics so I can finish this post before tomorrow:

-The Xbox does vertex processing on the GPU. The PS2 does it on the EE's VU1. So even if VU1 vertex processing was faster than the NV2a's (hint, it's NOT), that CPU power is not available to PS2 game code (as it's tied up doing vertex processing).

-No matter how you slice it, the EE comes up massively short compared to the Xbox's P3. The lack of a L2 on the PS2 is crippling. If you had actually developed for these systems, you wouldn't spout this garbage.

You, sir, are a tard.

Re:Cell- the first processor to run on hype

[be-fan]

I never said otherwise, you twit. Sony claimed that the EE would be twice as fast as a 733Mhz P3. It is at least twice as fast as a 733Mhz P3. The fact that the EE is more burdened in the PS2 really doesn't change that.

Re:Cell- the first processor to run on hype

[i41Overlord]

The EE *is* twice as fast as a 733MHz P3 No, it is not anywhere close to being twice as fast as a 733 mhz P3. I'm interested in where you heard that.

what's funny is..

[Feyr]

they say this cpu has 234 million transistors!

is it just me that has a bad memory, or is that around 5 times more than any other cpus on the market right now? this is just plain SICK

hope they got some pretty nifty technology to prevent this baby from going up in smoke, because i am NOT attaching a huge A/C to that thing! i find my p150 is noisy, and the p4 laptop unbearable. can't imagine what this will be like

Re:what's funny is..

[be-fan]

Prescott has 125M transistors, while the GeForce 6800 has 222M transistors. And on a tangent: this is typical Slashdot. IBM and Sony announce a 256 gigaflop chip, and Slashdotters' first reaction is to bitch about how hot and noisy it will be! Where are the real nerds in the audience?

Re:what's funny is..

[Bri3D]

Higher clock speed does not mean higher temps(that's only in Intel land). Your P150 probably produces a good bit as much heat as your P4, because at the time it was a fast CPU and used quite a bit of power. As CPUs get better the technology for less power(and usually thus less heat) should get better too, keeping the temprature about the same as the clock speed rises. Sadly that isn't the case with Intel's clock-speed-jacking powerhouse. So though the Cell will probably run pretty hot it was probably engineered not to run too hot.

Re:what's funny is..

[gkwok]

Yes, your memory concerning CPUs is serving you correctly.

For comparison purposes, an NVIDIA GeForce 6800 GPU has 222 million transistors.

Re:what's funny is..

[alienmole]

Higher clock speed does not mean higher temps(that's only in Intel land).

Actually, all else being equal, higher clock speed *does* mean higher temperatures, not just for Intel, but for everyone. IIRC, it's a linear relationship, too. However, as you say, there are some tricks, like reducing voltage, which can reduce power usage. There's a limit to that, though. The bottom line is that heat is a real limiting factor in speeding up computers by increasing clock speed.

Re:what's funny is..

[Bishop]

The Pentium 150 and 166 could run with a dinky little heat sink and the fan was optional if the case had a decent thermal design.

Re:what's funny is..

[Bishop]

If anyone cares the P200 (non-mmx) had a max power dissipation of 15.5watts, the P166 14.5watts, and the P150 was 11.6watts. See page 31. Other legacy spec sheets.

Re:what's funny is..

[Bri3D]

Actually, all else being equal, higher clock speed *does* mean higher temperatures, not just for Intel, but for everyone. IIRC, it's a linear relationship, too.

Not really. The clock speed is simply how fast a crystal vibrates. So in reality it is possible to build a CPU with high clockspeeds that runs really really slowly, for example one with a 6000 step pipeline. You could probably make one of these run cool too because there would be few transistors to deal with in the chip. The real relationship is between transistors and heat. The more transistors there are producing heat the hotter the chip is going to run.

Re:what's funny is..

[alienmole]

Actually, all else being equal, higher clock speed *does* mean higher temperatures, not just for Intel, but for everyone. IIRC, it's a linear relationship, too.

Not really.

Yes, really. It's basic physics - higher frequencies require higher energy, which means more power, which means more heat. The fact that laptops with power management slow their CPUs to save battery life ought to give you a clue: by saving power, they're also saving heat, and they do that by slowing down their CPU. Another place you see this effect is in CPU over-temperature safeguards - when the CPU temperature gets too high, the clock is slowed down to reduce the power consumption and therefore temperature.

There's some confirmation of this here: "Heat above spec can come from two areas: 1, Heat due to increased frequencies and 2, Heat due to increased voltage. Increasing bus speeds (frequencies) increases heat linearly"; and here: "There is a direct relationship between a processor's clock speed and the amount of power it consumes, and a similar relationship between power consumption and the amount of heat given off by a PC."

The clock speed is simply how fast a crystal vibrates.

Right, but that determines the frequency of the current pulses through the CPU and other chips. A higher frequency generates more energy, which creates more heat.

So in reality it is possible to build a CPU with high clockspeeds that runs really really slowly, for example one with a 6000 step pipeline.

You're correct that the frequency isn't a direct determinant of processor MIPS, for example, but that's besides the point. If the current flowing through the CPU is at a higher frequency, it generates more heat than would be the case at a lower frequency, simple as that.

You could probably make one of these run cool too because there would be few transistors to deal with in the chip. The real relationship is between transistors and heat. The more transistors there are producing heat the hotter the chip is going to run.

You're right, more transistors means hotter, but the temperature in each of those transistors is dependent on the frequency. If you figure out a way around that, you'd earn a Nobel prize and/or become a billionaire, because you'd be able to show Intel and AMD how to push their chips way past the 4GHz that's currently giving them a hard time (see the second article I linked above).

Re:what's funny is..

[aXis100]

Right, but that determines the frequency of the current pulses through the CPU and other chips. A higher frequency generates more energy, which creates more heat.

You're correct about the direct frequency heat relationship, but the above statement isnt very very clear.

In current semiconductors, a device (eg transistor) wastes very little energy in the ON (conducting) or OFF (non-conducting) states. Power = V * I, and in the conducting state, the voltage accross the junction is very low. In the non-conducting state, the current is very low.

Now, when you switch a semiconductor, there is a brief moment during transition when both the voltage and current are at intermediate values. The power consumption during this part of the transition spikes.

Faster clock speeds = more transitions = more power spikes and higher average power consumptions. It's that simple.

It follows that you can lower the consumtion by reducing the voltage, or making the transitions shorter (smaller, faster switching devices).

Re:what's funny is..

[alienmole]

Thanks for the clarification & explanation.

Re:what's funny is..

[Bri3D]

You are correct. What I was trying to say was that the relationship between the marketing department's clock speed number and the actual temprature and speed of the chip is NOT linear. The same chip with the same transistors at higher frequency will run hotter(parent's post in a nutshell).

Yes, but...

[Locke2005]

How much power does the chip draw? I'm used to heating my entire house with Intel processors; if this chip is significantly more efficient, then I'm going to get pretty cold next winter...

Re:Yes, but...

[rpdillon]

You jest, but the only room in my house where I can keep the themostat set at LOW all the time is in my computer dungeon - it's always toasty warm in there.

I run a three computers, but none P4 - Athlon 64, Athlon XP 2600+ and a Intel Pentium M (laptop).

On topic, I really want to get a box powered by this and run Linux on it...that'd be sweet. And to the guy who said there was no point in having a 4GHz Linux box because there were no "real" games: I hope you jest. There are *few* games, but they are *real* for sure: Doom3 and UT2004 come to mind, as well as Quake 3 and Neverwinter Nights. All can take more CPU cycles, expecially UT and Doom.

And heck, mprime could always use some extra cycles. =)

Re:Yes, but...

[Lord Crosis]

I don't know where this analyst gets his information, but:

"Cell will probably consume around 30 watts of power, similar to the Emotion Engine processor in the PlayStation 2 console, said Peter Glaskowsky, a technical analyst with The Envisioneering Group in Seaford, New York. This is also similar to the power consumption of Intel Corp.'s Pentium M processor."

This from the MacCentral article on Cell.

-=(Lord Crosis)=-

Ah!

[be-fan]

The math makes sense now. 8 APUs, 2 pipelines per APU, 4 operations (SIMD) per clock, at 4GHz. That gives 256 gigaflops!

Re:Ah!

[djward]

And the power supply it needs outputs 1.21 Gigawatts.

Unlikely.

[katharsis83]

"plan to recoup their R&D expenses largely from other consumer multimedia devices and NOT from selling Cell processors or Cell processor based computers, meaning they will be surprisingly inexpensive. Yeah!!!"

However, from the press release:
Prototype die size of 221mm2

When it comes to chip manufacturing, the cost of a chip is basically a direct function of the area. A 221 mm^2 chip size is pretty damn big; this thing isn't going to be cheap. Even considering IBM's extensive fabrication experience, Sony will probably have to sell this at a significant loss to make the PS3 palatable to gamers.

Granted, this is a prototype, so they can probably shrink it further by production, but it still won't be something cheap. Don't count on being able to buy these cheaply to make your own parallel supercomputer.

Re:Unlikely.

[drinkypoo]

Nothing, except that they're going to be soldered down and it's going to be a gigantic pain in the ass.

Re:Unlikely.

[Locke2005]

Pardon my naivete, but if they go from a 90nm process to a 60nm process for actual production, shouldn't that make this a ~150mm~2 chip size? And isn't cost really a function of yields (which go down with increasing area), not "a direct function of area"? We have yet to see what kind of yields IBM can achieve with the final process -- until then, nobody can really speculate on cost.

Re:Unlikely.

[SQL Error]

No, when they go from 90nm to 65nm, the die size will drop to about 115 mm^2 - about half the size. (Remember, we're measuring area here, and the process size is a linear measure, so you have to square the ratio.)

65nm is the target process for the Cell. The 90nm chips will be for development workstations and such.

And yes, cost is a function of yield, but yield is largely a function of area. The smaller the chip, the less likely there will be a defect.

Re:Unlikely.

[Locke2005]

I thought I did square the ratio! But I must have messed up my math somewhere, because you are right, the correct size should be ~115.27 mm^2 for 65nm process!

It must be said

[gkwok]

If so a cluster could prove quite formidable indeed!

Slashdot translation: Imagine a beo-- never mind.

Re:About time.

[kamapuaa]

Now my dreams will have a visual to go with them...

My dreams already do.

Re:Cell Processor enhances cross-platform software

[Locke2005]

Sure, you'll be able to play doom III on your cell phone! But you'll have to wear oven mitts to keep from burning your hands, and the battery life will be about 5 minutes!

Supposedly, since the Cell is designed for distributed processing, it will be usefull in cellphones... but I'd rather have a cellphone that I can reliably talk on, rather than one that a cracker halfway around the world can take over and use as a spambot! They seem to be making a lot of promises about the Cell hardware that can only be realized in software -- software that hasn't been written yet!

Re:Xbox

[SteveXE]

Dont forget Sony CANT block MS from using Cell in Xbox 2, all IBM has to do is change the layout and sell it to MS and you now have 2 consoles with Cell chipsets in them. We already know Xbox is using a Power based CPU, we just dont know why...and Cell is Power based.

cell processor

[jotux]

so does this mean XIII will run even better?

Dont jump the gun...

[SteveXE]

Before everyone starts having digital orgasms over this chip we should take a sit back and watch approach. This all sounds amazing on paper as did PS2 and look how that turned out in terms of hardware power.

We have no idea if developers will be able to easily adapt and get any real performance out of this thing above and beyond what they get from CPU's now. Almost nobody uses the vector units in PS2, who says they will start now? In terms of just gaming I wanna see some games and examples of this thing running in real time before i start taking my wallet out of my pocket, Sony burned me last time with underpowered bug ridden hardware, ill be damned if i let them do it again.

Re:Dont jump the gun...

[Ziviyr]

I want to see examples of this thing rendering graphics with filtered textures.

PS2 and PS1 are pretty sad there.

Re:Dont jump the gun...

[be-fan]

Cell != PS3. It will be the PS3 processor, but this article is about Cell, not the PS3. So texture filtering really has nothing to do with this article, because that's a function of the rasterizer Sony puts in the PS3.

Re:Dont jump the gun...

[Ziviyr]

I dont recall seeing much more than the EE on the PS2 for graphics. I think Ive read enough articles about it, either the 3D hardware on it isnt noteworthy, or the EE does the bulk of the work.

I expect similar on the PS3 for some reason...

Re:ANY ___FACTS___ AT ALL??

[be-fan]

Did you RTFA? From the second article:

Die size: 221mm^2
Transistor count: 234m
SPE Size: 2.5x5.81mm
SPE Interconnect: 4x128bit ring bus
SPE local memory: 256KB
SPE decode rate: 2 insns/cycle
SPE resources: 7 execution units (unspecified type)

They also mention the core voltage of the CPU (1.3V), the fact that the memory has been tested to 5.4GHz, detail the temperature monitoring scheme, and the fact that the SPEs are in-order chips. This is all new information.

Re:I wonder....

[cosmo7]

I wonder what M$ has to beat back a server processor with essentially hyper threading, running at 4.6 ghz, attached to 8 vector processors, each with a lot of registers and cache, which are using extremely fast memory, that can connect to other, similar processors nearby.

Microsoft has consistently overwhelmed the fastest processors on the market and I am confident that with the right bloatware they will continue to do so.

Your wish is granted

[nacturation]

Here you go: http://it.slashdot.org/article.pl?sid=05/02/07/231 0237

Re:I wonder....

[ryanmfw]

I meant hardware wise for the XBox 2! But you are most certainly correct on M$ bloatware!

Power and heat

[Sgt_Astro]

It runs at 1.3v as I've read. They say it's temperature is between 40C and 52C. Nobody is going to get cold this winter, but you aren't going to need water cooling either.

The reason this chip isn't as hot as the sun is because its being produced by a smaller nm process than current processors. The smaller the nm process the smaller the transistors and therefore less power and heat are required for the same clock speed. AMD is still using a 130nm process to produce most of its processors with a 90nm factory opening soon to produce a new line of chips. These Cell processors are going to be produced using a 60-65nm process in factories that are currently under construction by IBM, Sony and Toshiba.

The factories to mass produce these things aren't even finished yet, so don't get your hopes of seeing one up until mid summer 2006.

P4 processors have 125 million Transistors in them now and use a 90nm process. If they double that number (ie. 2 cores anyone), they will have 250 million transistors, 15 million more than the 235 million purported to be on the cell.

Re:Cell Processor enhances cross-platform software

[action789]

I'd be far more worried about malcode (virii, worms, etc) taking hold in the cross-platform evironment the parent mentions than worrying about games.

Re:Xbox

[Froggert]

It's highly unlikely that this would happen. The most obvious argument against this that I see is that although the CPU cores may share some common ancestry, what really drives console games is the fact that you can squeeze every tiny bit of performance out of the machine by coding directly to the hardware, i.e. video and sound. If the memory architecture, etc. is different, then the perf is going to be totally different on two different platforms. Just my 2 cents. Cheers!

Re:I wonder....

[WindowPane]

Please mod parent +3.

Super TVs

[J Mack Daddy]

From TFA:

"Cell-based products will be used in devices ranging from digital televisions to home servers to supercomputers, they said."

Sweet, I've always wanted my TV to double as a supercomputer. Maybe then it can tell me if the weather guy is yanking my chain.

Missing the point

[egrinake]

There seems to be alot of confusion surrounding the Cell chip. This is not "just another processor", and it certainly has little to do with clock frequencies - the Cell is a whole new architecture, which might just be a glimpse into the future of computing.

To begin with, it might be useful with some background on the ps2 architecture - there are a couple of really great in-depth articles at Ars Technica; Sound and Vision: A Technical Overview of the Emotion Engine and The PlayStation2 vs. the PC: a system-level comparison of two 3D platforms.

What made the ps2 so awesome was that it was custom-built specifically for multimedia-processing, which requires completely different processing environments than general-purpose computing. Normal PCs are made for computing where you have a large number of instructions working on a small data-set (such as a spreadsheet) - this requires large data-caches close to the CPU, while instructions are streamed continually from RAM. Media-processing is the other way around; you have "simple" operations (like doing the calculations for a single pixel), which are run on a large set of data - so you wouldn't really need any data-caches. The ps2 did exactly this; it removed almost all the caches (only a few tiny ones were left), but it had a totally insane bus bandwidth. To borrow an analogy from the mentioned Ars Technica article:

"Here's a goofy example to help you visualize what I'm talking about: imagine a series of large buckets, connected by pipes to a main tank, with a cow lapping water out of each bucket. Since cows don't drink too fast, the pipes don't have to be too large to keep the buckets full and the cows happy. Now imagine that same setup, except with elephants on the other end instead of cows. The elephants are sucking water out so fast that you've got to do something drastic to keep them happy. One option would be to enlarge the pipes just a little (*cough* AGP *cough*), and stick insanely large buckets on the ends of them (*cough* 64MB GeForce *cough*). You then fill the buckets up to the top every morning, leave the water on all day, and pray to God that the elephants don't get too thirsty. This only works to a certain extent though, because a really thirsty elephant would still end up draining the bucket faster than you can fill it. And what happens when the elephants have kids, and the kids are even thirstier? You're only delaying the inevitable with this solution, because the problem isn't with the buckets, it's with the pipes (assuming an infinite supply of water). A better approach would be to just ditch the buckets altogether and make the pipes really, really large. You'd also want to stick some pans on the ends of the pipes as a place to collect the water before it gets consumed, but the pans don't have to be that big because the water isn't staying in them very long."

So, what does this have to do with the Cell? The Cell takes this concept even further. Cell systems are made up of multiple processors, called APUs (Attached Processing Units), which are connected using an insanely fast data bus. Each APU can be programmed to handle one specific task, and then pass the data on to the next APU for a different task. By doing this, you can just put in more processors to increase the throughput of the system. This works especially good for multimedia processing, which can be pipelined like this pretty easily. Here are a couple of snippets from the Wikipedia entry:

"While the Cell chip can have a number of different configurations, the workstation and PlayStation 3 version of Cell consists of one "Processing Element" ("PE"), and eight "Attached Processing Units" ("APU"). The PE is based on the POWER Architecture, basis of their existing POWER line and related to the PowerPC used by Apple

Re:Missing the point

[X]

I'm not sure that this makes for the "extreme goodness" that people are envisioning. It should be truly awesome for SIMD type operations, such as video encoding/decoding and 3D rendering. But I don't see much of a break through here in more general compute tasks. If you've worked with an Itanium, you know the parallelism limits you hit with VLIW instruction sets and having a bunch of VLIW processing units is just going to make it worse.

All in all, this thing strikes me as more of a next-generation DSP rather than a next-generation CPU, with a lot of hype thrown in (btw, the are apparently now called "synergistic processor elements" instead of "attached processing units" ;-).

Re:Missing the point

[Spy Hunter]

So what I'm wondering is: will Cell make 3D accelerators obsolete? For some time graphics cards have been moving away from single-purpose graphics circuits toward general vector processing. CPUs are coming the other direction, from fully general instruction sets to specialized vector processing instructions. Cell seems to be the logical merger between a CPU and a graphics card. Will Cell workstations have the raw vector-processing power to make a 3D accelerator card redundant? Or will specialized graphics cards still be able to have a significant edge in rasterizing textured polygons?

Re:Missing the point

[zymano]

more usage = better compilers.

We need to get away from the intel isa .

We shouldn't need to dumb down processors just for people to use word processors.

Re:Missing the point

[carlmenezes]

Well, from what I've read so far, this is what I understand :

Everyday applications require lots of operations on small sets of data. Multi-media apps require simple operations on large sets of data.

So, theoretically, couldn't each unit be setup to provide a subset of those "many operations" and then be used on a small data set?

Or, what about having a cell "co-processor" that takes care of stuff like anti-aliasing, drop shadows, and other graphics stuff that is common in everyday use? Would this eliminate the need for a graphics card? I mean, if we put a cell processor on the motherboard next to the CPU, we should be able to have a very capable video card like unit, without something like AGP choking it, right?

Re:Missing the point

[kpp_kpp]

back in the 80s or early 90s i saw an article by someone which the gist was "processors will have to change architectures just like trains had to switch from steam engine to diesel"....

you could theoretically build a single steam engine that was more powerful than a single diesel engine but you could not hook two steam engines together.

looks like we've finally just about hit the limit as to where a single processor can take us in regards to speed... from now on it will be chip architectures that can easily be chained together that will have the advantage.

Re:Missing the point

[X]

So, theoretically, couldn't each unit be setup to provide a subset of those "many operations" and then be used on a small data set?

This already happens today inside of a CPU. There are limits to what you can do because of depedencies and branching. The Itanium pushes this at least as far as anything I've seen about the cell processor.

Yes, you could use the cell as a co-processor for graphics and other operations. You may recall that IBM made a DSP called the MWave a while ago that worked like this. In the end, it didn't turn out to be that big a win over using specialised chips.

Just because you are using a "co-processor" instead of a "GPU" doesn't change the bandwidth bottenecks in the system. In fact, current GPU's basically never saturate their AGP or PCI express interfaces, but instead saturate internal limits which the graphics card makers do their best to alleviate. The GPU guys work really hard to maximize the performance of these systems, and it's a very competitive market. The cell might be somewhat more efficient at this, but the architecture isn't going to wipe out the competition. The only thing that might make the cell really different are economies of scale.

Re:Missing the point

[X]

more usage = better compilers

Eh?

We need to get away from the intel isa

There is no "intel isa". They ship processors with a variety of different ISA's. History has taught us that ISA's in the end don't make nearly as much of a difference as other, far more important factors.

We shouldn't need to dumb down processors just for people to use word processors.

Again, you aren't making any sense. Are you suggesting somehow that existing CPU's are "dumbed down" because the designers are concerned about how well they'll run word processors?

Re:Missing the point

[iwadasn]

The PS2 was awesome? Brother, you've been misinformed. The PS2 was supposed to render hollywood quality graphics in realtime, and cook you breakfast as well. When it came out it was competitive with a decent PC for graphics, though it got blown out everywhere else (ram capacity, etc.....). The XBox eats it alive. If you see the two side by side it's hard to deny that the XBox graphics are hugely better.

The point is that Sony has a nice long track record of designing silicon with huge theoretical potential. Unfortunately, the designs often turn out to be almost impossible to program decently. The Cell might be different, but I kindof doubt it. It'll be competing with multicore vector enabled PPC and x86 processors when it finally comes out, and it'll be harder to program than any of these designs. Its theoretical performance might be higher, but I'm not sure anyone will be able to actually make it run faster than a modern PC when it's released. We were supposed to have supercomputers made from PS2 emotion engines by now, where are they?

Re:Missing the point

[egrinake]

Well, even though the ps2 and pc-based systems had similar performance, I still think that the ps2 architecture was a step in the right direction. Even though you are able to do alot of cool stuff with a pc-architecture, I'd much rather see further development on media-tailored systems than trying to teach an old dog new tricks. Current PCs are trying to be alot of different things at the same time, and I think it's pretty obvious that specialized systems will be much better in the long run.

But then again, I'm not an expert on any of this, I'm basically just parroting information I've read elsewhere. I don't know anything about the xbox 2, so I can't comment on their relative merits, but the Cell seems like it at least should be able to do a much better job at media-processing than current PC-style architectures.

Re:Missing the point

[zymano]

Our supercomputers are definitely dumbed down.

Parallel processing is the future. Accept it.

Re:ANY ___FACTS___ AT ALL??

[gl4ss]

if by second article you mean the http://www.blachford.info/computer/Cells/Cell1.htm l it's not really credible at all.

it's not real facts about what the _real_ cell will be able to do when it's out. it's just steam in the hype machine guessworking out something along the lines of "oh boy oh boy oh boy it's gonna be sooo goooood and this |-------------------| fast".

it's not out yet - there's no real facts about what it can do, nor is there real facts about what clockspeeds it will run and what kind of integer/traditional _real_ performance you can get out of it.

so these 'articles' are quite useless. they're 'reviews' done with marketing material, some of which is starting to get quite old marketing material too and was probably optimistic too, of course, because you can promise *anything* if it's two years into the future when you'll have to actually deliver.

the biggest point however is that sony overhyped it's last consoles cpu just as much and failed to deliver utterly.

Terminators

[EnsilZah]

Does that make terminators the aimbots?

I did, I'm still confused

[mcc]

Hi.. so I am trying to take all of this in. Please help me understand. I am a Mac programmer and I have a relatively good understanding of what's happening in the PowerPC. I'm a bit confused about some elements of this Cell thingy though.

• So the CPU is just a normal POWER, right?
  
• But, from the article: "Along side these is a 64-bit Power processor capable of running two threads." "Capable of running two threads"? Is this the same as hyperthreading in the intel processor? I did not know IBM was working on their own implementation of that, have multithreading POWER CPUs been used in a product yet?
  
• It says t*he vector unit is a "VMX", well, that's just the same thing as Velocity Engine / Altivec that Apple uses, it's just a different brandname, right? And that'll be just part of the POWER, like the Altivec unit would be on a PPC?
  
• The SPEs/APUs/"stream processors" are in particular what's confusing me just a little. I can think of lots of circumstances for which these things would be useful. But what I don't get is why, if you have these things, you still need the VMX. For what purposes is the VMX more suited? Has it got better throughput for the applications to which it is suited? Does it work better with the main CPU than the SPEs? Or is the idea here just that you don't have to keep the SPEs busy doing stuff that a normal VMX could handle?
  
• Here's my big concern: on the Mac, the big problem with altivec has been keeping the altivec units fed. A lot of the time the altivec isn't getting used for anything, and even when it is, prior to the G5 there were serious problems with pushing enough data through the bus to keep the Altivec constantly busy. Will the SPEs have this same starvation problem? From the article: Connecting up the processing units is the element interconnect bus (EIB), comprising four 128-bit rings and a 64-bit tag running at half the processor clock. The busses connect to the SPEs through local memory, 256kbyte for each SPE. .. What is this telling us? That each of the SPEs has 256k of private memory to work with? Can SPEs freely read other SPEs "local memory", or only their own? And who fills up this memory initially, and who deals with it once it's done? The main processor? I.E., do the SPEs have access to main or video memory or other hardware, or do they ever require for the CPU to shuttle data to keep them fed? It appears from the original patent overview that the SPEs can talk directly to the memory controller, so if I'm reading that right then that's good-- that seems to make them qualify as actual processing units and not just coprocessors like the VMX is. But then the article seems to be saying the is SPE access to memory is limited-- i.e. it can only be done in block load/stores. Well, are there other limitations, can the APUs talk to (for example) video memory?
  
• But, crucially, who loads the instructions for all of this?? If we've got a CPU that can be running two threads, and 8 little APU/SPEs that are each effectively running as their own processor, and all of this is sharing one memory bus... that's like effectively ten instruction streams to be reading at the same time. Is that going to be a problem? Do each of the 8 SPEs actually independently load their own instruction streams? Or is the idea that they partially use that 256k "local memory" as effectively an instruction cache?

If you can help clarify some of this for me, thanks.

Re:I did, I'm still confused

[pohl]

I'm looking at this and thinking some of the same things (specifically, if we have these SPEs what do we need the Altivec unit for?

But then I thought about all of that cool Core/Image/CoreVideo stuff that was demonstrated in a Stevnote about a year ago, where there were all of these fancy realtime imaging and video filters that had a very interesting implementation: the filtert source code is fed into a special compiler that offloads the filter (or chain of filters) onto the GPU of your video card. I'm tempted to think that all of that work of using the GPU in this way was just laying the groundwork for this anticipated CPU. In the very least, one would think that this processor could take the place of that aspect of the GPU, and then the video subsystem potentially gets a lot simpler (just add framebuffer) which could open the door for squeezing more into small form-factor machines.

Excuse, me, I'm having a bit of a koolaid rush.

Re:I did, I'm still confused

[be-fan]

So the CPU is just a normal POWER, right?

No. Each Cell has one main (controller) CPU called a PU, and up to 8 seperate vector CPUs called SPEs. The main CPU is a regular 64-bit POWER processor (with SMT --- IBM's equivalent of hyperthreading), while the APUs are very simple processors with a lot of execution resources and insane bandwidth. Such processors are known as "stream processors" in the literature, because they are designed to handle streams of data.

it's just a different brandname, right?

Yes, "AltiVec" (like "G5") is an Apple/Motorola trademark, so IBM can't use it. And you're right, the AltiVec unit is on the PU.

For what purposes is the VMX more suited?

It's there most likely because if you're running some code that isn't suitable for the SPEs, but does need to do vector computations, you don't have to send it off to the SPEs.

Will the SPEs have this same starvation problem?

Potentially, but probably not. Altivec on the G4 was starved because the G4's bus was exceedingly slow. The SPEs are supposed to be on a shared 128GB/sec internal bus, and the Cell has 100GB/sec of bandwidth to main memory.

That each of the SPEs has 256k of private memory to work with?

Yes. In the Cell model, you design your code in "cells". A cell is a clump of code and data that's copied to the SPE's local memory. The code then runs, streaming in additional data from memory, and using the local memory as a workspace.

Can SPEs freely read other SPEs "local memory", or only their own? And who fills up this memory initially, and who deals with it once it's done?

The SPEs local memories are not connected to each other, so each SPE can only read from its own local memory. The memory is filled up by the PU, when a Cell is loaded onto the SPE. The SPE then runs autonomously, and when it finishes, sends the results back to the PU via main memory.

I.E., do the SPEs have access to main or video memory or other hardware, or do they ever require for the CPU to shuttle data to keep them fed?

The SPEs and the PU all talk to a single DMAC, which has access to main memory.

But then the article seems to be saying the is SPE access to memory is limited-- i.e. it can only be done in block load/stores.

Yes. The DMAC, actually, can only read/write in 1024-bit blocks. This isn't really a big deal if you think about it. When a regular CPU reads a memory address, it doesn't read a byte at a time. It loads a whole cacheline at a time. So a P4, for example, usually reads a 128-byte (1024-bit) block at a time from memory anyway.

Do each of the 8 SPEs actually independently load their own instruction streams?

Yes. All the processor units run seperate instruction streams. Each "software cell" runs in its own thread, if you will.

Re:I did, I'm still confused

[temojen]

I don't work for IBM, so I only know what I've read about the Cell, and other designs.

So the CPU is just a normal POWER, right?

Sounds like it.

But, from the article: "Along side these is a 64-bit Power processor capable of running two threads." "Capable of running two threads"? Is this the same as hyperthreading in the intel processor? I did not know IBM was working on their own implementation of that, have multithreading POWER CPUs been used in a product yet?

Sounds like it is the same as HT. I don't know if they have any in production yet.

It says t*he vector unit is a "VMX", well, that's just the same thing as Velocity Engine / Altivec that Apple uses, it's just a different brandname, right? And that'll be just part of the POWER, like the Altivec unit would be on a PPC?

Altivec is a trademark owned by Mototolla, velocity engine is a trademark owned by Apple, and VMX is a trademark owned by IBM. They all refer to the same thing.

The SPEs/APUs/"stream processors" are in particular what's confusing me just a little. I can think of lots of circumstances for which these things would be useful. But what I don't get is why, if you have these things, you still need the VMX. For what purposes is the VMX more suited? Has it got better throughput for the applications to which it is suited? Does it work better with the main CPU than the SPEs? Or is the idea here just that you don't have to keep the SPEs busy doing stuff that a normal VMX could handle?

Stream processors tend to have a lot better processing throughput than vector units as they do all the manipulations on each data element before moving on to the next one. Vector units may zip through the same array several times, requireing multiple loads of the same data. Also, stream processors are optimized for situations where each set of data can be manipulated independently from the others, so you can use several of them at the same time.

Here's my big concern: on the Mac, the big problem with altivec has been keeping the altivec units fed. ... Will the SPEs have this same starvation problem?

Probably not. Each SPE has it's own registers (lost of them), and private memory (probably not readable by the other SPEs), and it's own instruction cache. They're designed to run a tight inner loop (called a kernel) over their chunks of data, performing all required operations before storing the result and moving on to the next chunk of data. See this overview of stream processing.

But, crucially, who loads the instructions for all of this?? If we've got a CPU that can be running two threads, and 8 little APU/SPEs that are each effectively running as their own processor, and all of this is sharing one memory bus... that's like effectively ten instruction streams to be reading at the same time. Is that going to be a problem? Do each of the 8 SPEs actually independently load their own instruction streams? Or is the idea that they partially use that 256k "local memory" as effectively an instruction cache?

Probably the CPU loads the kernel into the SPE instruction cache.

Stream processing was designed to compensate for the memory bandwidth issues associated with Vector Processing, and to take advantage of ALU Paralellism.

According to that overview:

There is a class of applications that lend themselves to being streamed processed. They have several defining characteristics. First, they provide an opportunity to exploit parallelism. Secondly, they have little or no data re-use, which makes using a conventional cache expensive. Finally, these applications have high computational intensity, which enables a high ratio of computation to data.

Re:I did, I'm still confused

[mcc]

Thank you for your responses!

Re:I did, I'm still confused

[temojen]

If you want to try out stream processing now (without having to buy a supercomputer), check out BrookGPU.

Re:I did, I'm still confused

[Woody77]

Yes. In the Cell model, you design your code in "cells". A cell is a clump of code and data that's copied to the SPE's local memory. The code then runs, streaming in additional data from memory, and using the local memory as a workspace.

This is, essentially, reconfigurable computing. Something that's never quite made it to market, by because the time that the FPGAs-type hardware was cheap enough to do it, general purpose cpus were faster again.

But this looks like an excellent cross-breed. Instead of flashing an FPGA/DSP and throwing data at it, instead just download some microcode into the instruction ram of the cell, and then tell it where to find the input and output streams, and have at it.

Depending on the size of the stream execution code in the cell, you could really do some amazing things in applcation servers with something like this (like using a series of cells to do SAX and XSLT transforms on data in realtime).

The key to usefullness for this, though, is management interface for the stream processors. Are they I/O entities? Are they managed by the compiler? the OS? How fast do they context switch, and if slow, how do you manage the proper allocation of them?

Watch Windows 2010 or Gnome/KDE use half of them for rendering the UI....

Mod parent UP please

[Burz]

This is about IBM, Apple and Sony ganging up on Microsoft in a BIG WAY.

It certainly does appear that Mac OS X could be the primary beneficiary of the new CELL processor.

Re:Mod parent UP please

[Alan Partridge]

Wait and see, there's no reason whatsoever to believe that Apple will use Cell yet - though there's plenty of good arguments to be made for them doing so.

They already tried this on PS2!

[jbischof]

They already tried this for the PS2.

Re:They already tried this on PS2!

[cgenman]

It's funny they say that "it could provide the processing power for the PlayStation 2 to challenge cheap PCs as the entry-level device of choice for home access to the Web."

Since when is a console a replacement computer? And for that matter, why would you need a hot $hit console to surf the web? The processing power needed to bump off cheap PC's? Get real. You could surf the web on a Saturn. There was probably enough oomph in the Super NES to do it.

Why must journalists always find something for game consoles to do other than play games? Games are a worthy enough pursuit to have their own hardware. You don't see the tech reporters questioning DVD player makers how they plan to engage in more worthy pursuits like spreadsheet operations and surfing whitehouse.org. Why must consoles be more than consoles?

Re:Cell Processor enhances cross-platform software

[SteveXE]

the whole point of the cell in a cell phone would be not needing a gpu or large cpu to run graphic intensive games. with cell we should be able to get at LEAST ps2 level graphics in a small cell phone.

Re:Xbox

[SteveXE]

Doubtfull since im sure IBM did most of the hardware work themselves. Its 3 companies, if anything happens it will need majority aproval...i would assume, in that case IBM, MS, and Toshiba all have "good" buisness relationships at the moment, and having MS on board would only benefit IBM and Toshiba with adoption time. More companies who license the tech=more money and no matter how big Sony's grudge may grow it wont out vote the other 2

Re:SPU's are really FPU's?

[jonabbey]

The SPU's are something like general purpose processors, from the earlier reports, except they do their general purpose work on chunks (cells) of code and data that are sent to them. Each SPU has its own high-speed dedicated attached memory.

The SPU's are optimized for streaming vector operations, but it seems they have their own branch and conditional instructions as well.

Almost nobody uses the vector units on PS2?

[jonabbey]

Then how are people writing the kind of games that are coming out for it now?

Re:Xbox

[steeviant]

pff... It better be more advanced than that. I've got a shotgun that can do what you mention.

Cell == Sony Exclusivity or Desktop Use?

[failedlogic]

Just wondering if Sony is keeping exclusivity for this technology? Or, has IBM developed it for them and can use it on (possibly) desktop PCs but not in other consoles?

Re:Cell == Sony Exclusivity or Desktop Use?

[be-fan]

IBM has developed for a wide range of uses. They are talking about a digital content creation (DCC) workstation with Cell, as well as rackmountable units for scientific computing.

The next Xbox will not have a Cell processor

[Mal Reynolds]

You're making a common, but incorrect assumption. Yes, both Microsoft's and Sony's new consoles will be using IBM processors. But there is no way in hell Microsoft is getting the Cell processor for the next Xbox.
The new Xbox is rumored to be using dual-core Power PC chips. Nothing very new there and not very different from what one can buy in a Mac today.
To see why Microsoft won't be getting the cell, one only has to consider the huge level of investment Sony has made to the initiatiave. Not only has Sony been involved in the development of Cell for a number of years, they're even building their own multi-billion dollar Cell chip manufacturing plant.
Would Sony become this involved in development and production of a revolutionary processor just to let their leading console rival use it right out of the gate? I could almost guarantee you MS is contractually locked out of the Cell for a few years yet.
The final straw is that the next Xbox is probably going to be shiping before Cell chips are even in mass production. I guess you might see a Cell in Xbox 3, look for that around 2009.

Re:The next Xbox will not have a Cell processor

[fyngyrz]

But there is no way in hell Microsoft is getting the Cell processor for the next Xbox.

Oh, I don't know. Microsoft could just up and buy IBM. They might have trouble buying Sony, given that it's not a US company and that Japan has been known to be fiercely protective of its corporate citizens, but IBM? I doubt that'd be a serious problem. If it breaks laws, they'd just buy the politicians -- and it wouldn't be the first time.

Never underestimate the power of cash to alter the world you live in.

Re:The next Xbox will not have a Cell processor

[fyngyrz]

Sure do. Looked at MS lately? Looked at Billy-boys cash value? Do you realize that you don't have to buy everything, just a majority of the stock, to control the company? Can you say hostile takeover with a Boston lawyer accent? :-)

I don't think it's even a question of if MS could; only a question of if they would want to. If the Ps3 can truly kick the XB2's butt (and that's as much a matter of software as it is of fabulous hardware) then just perhaps they would.

MS might just throw the current XB2 design in the toilet and sell software for the PS3, too. There's nothing quite so predictable, yet as blatently unexpected, as tomorrow's news. I saw that coming! Yeah. Sure I did. :-)

What's that you say? MS wouldn't invest in supporting a hardware architecture and then just let it go? Er... NT Alpha. NT PPC. NT MIPS. Here yesterday, gone... yeah, yesterday, that's the ticket. XB2? Not here today... and not here tomorrow, either?

I bet there are a lot of XB2 engineers crying in their beer tonight. It might not be as bad as it seems, but... it sure seems pretty bad. Beer with salty tears. Yech. :-)

Re:ANY ___FACTS___ AT ALL??

[be-fan]

Sorry, I mean the second article out of the two new ones. Blachford's is an older one. Anyhow, I don't see how these aren't real facts. They cite some very important points about the architecture, and to someone who knows how CPUs work, they are quite useful in predicting the performance of the thing. In particular, the fact that the SPEs are in-order tells you a whole lot about how important instruction scheduling is going to be.

With regards to the EE, it wasn't a failure at all. The EE delivered exactly what it promised --- 6.2 gigaflops at 300Mhz. The only people who saw other promises were the ones who bought into the crap they read on CNET or some other piece-of-shit site, and people who have sense have learned to filter those sources out instinctively. In nearly example of supposed "hype" I've seen about the EE, it was the result of a journalist trying to "put the numbers in context" for their clueless readers.

Intel not impressed

[vandan]

We are very reluctant to adopt architectures like this because they take compatibility and throw it out the window.

You mean like the Itanic? Shoe's on the other foot now, eh?

Re:Intel not impressed

[Ohreally_factor]

"Asymmetric architectures confuse and frighten me."

-- Intel's Caveman Spokesman

Re:Intel not impressed

[MeridianOnTheLake]

You mean like the Itanic?
Not really. The Itanic is targetted at the high-cost, low-volume market segment (that's how it turned out, anyway). The cell, however, should be produced in much higher volumes by virtue of it being used in games consoles and hence should hopefully be available at a much lower price-point. A lower price means more people will experiment with it, more Linux users will buy it, leading possibly to a positive feedback cycle where the chip gets cheaper and cheaper, or maybe faster and faster, and hence used more and more.

It's not fast enough for me

[xtremee]

I can't wait to put my hands on this baby, buy a really BIG BIG cooler and overclock it!

Sure Sony could keep the Cell from Xbox

[Mal Reynolds]

Of course Sony could block Microsoft from using the cell in the next Xbox.
And no, changing the layout would just bring on a lot of lawsuits. Furthermore, some of the IP on that chip is purportedly owned by Sony. Even with a new layout, Sony would have to permit the sale, something they are not about to do. Playstation is Sony's largest net revenue source.
As for prohibiting MS use of the Cell, Sony certainly has a lot of contracts with IBM regarding Cell. And if one of those contracts prohibited sale of the cell to other console manufacturers, then IBM would be contractually prohibited from selling those chips to Microsoft.
Far more common outright prohibitions are contracts allowing exclusive time periods. For instance, Sony having a 1 year exclusive on the newest revisions of any Cell processor. Considering how infrequently consoles are released, even that relatively minor 1-year exclusivity would forever lock Microsoft out of the best Cell processors. Exclusive time period contracts like that are very common, and there is absolutely nothing illegal about them.
But the biggest reason the next Xbox won't have Cell is because the next Xbox is scheduled to ship before the Cell will be available in quantity. Maybe the Xbox After next will have a cell, but that's easily 4 years from now.

CELL Processor and Apple Speculation :: GCC?

[tyrione]

Unless Apple suddenly releases GCC support for the CELL Processor I doubt we'll see this for a Mac any time soon.

Re:CELL Processor and Apple Speculation :: GCC?

[Lord Crosis]

An excerpt from MacCentral's article on Cell:

"Cell is primarily designed for digital-home applications, but the chip can be used with a wide variety of programming models and could conceivably wind up in any number of systems, Kahle said. The companies are working with open-source compiler developers to create software development tools for programmers, he said."

While this isn't the same as Apple releasing GCC contributions, what open-source compiler could they realistically be talking about besides GCC? At the very least this seems to support the idea that Cell could be an incredible Linux platform, but ultimately if Cell lives up to 1/10th of the hype that was released today, and is indeed fully compatible with the POWER instruction set (implied but not verified), It sounds like Apple would be insane not to use it.

-=(Lord Crosis)=-

context switching

[ArbitraryConstant]

As I understand it, the APUs can act semi-independantly but the controlling processor has overall control. If that's true, if the processor wanted to (say) switch to some other process would it have to save all that state to somewhere else before continuing, just as standard processors do now?

As all the APUs have lots of big registers and significant amounts of private memory, wouldn't that be painful?

Re:context switching

[Tap-Sa]

As all the APUs have lots of big registers and significant amounts of private memory, wouldn't that be painful?

With 100Gbyte/sec memory bandwith, no.

Re:context switching

[ArbitraryConstant]

Isn't that a security risk?

Re:context switching

[be-fan]

The SPEs don't context switch. They process a single APUlet until it signals conclusion, and then stops until another APUlet is loaded. I presume you could interrupt it and load a different APUlet, but the APUs aren't cores that you can schedule threads on.

Re:context switching

[ArbitraryConstant]

"APUs aren't cores that you can schedule threads on."

Right, but the controlling processor can be scheduled and it can context switch. How does it handle the SPEs when it does? There must be some mechanism to keep the SPEs from letting one thread/process's APUlets from bleeding into another's.

Re:context switching

[be-fan]

The OS has an APUlet batch scheduler. You submit an APUlet to the scheduler, along with its timing requirements (for time-sensitive jobs), and it will set up the APU, and send the results back when its done.

Re:I wonder....

[ciroknight]

Like, total duh dude. Haven't you heard of Longhorn?????

ipod in the ass, eh?

[I judge you]

It is highly likely that Sony will create a custom version of the PS3 hardare for its dev kit (as it did with PSX and PS2). If for some strange reason it decides to use an off the shelf system for an early dev kit, it would likely be the CELL-BASED WORKSTATIONS SOON AVAILABLE FROM IBM.

I diagnose you with incurable stupidity. You can go now.

Shoo!

All of this...

[eremitic]

and I bet it still wouldn't be able to run Longhorn.

Re:Cell Processor enhances cross-platform software

[diablomonic]

I dont think anyone is realising the true potential of these chips, all they are seing is 4.6 ghz power chip and forgetting the multiple stream processors.
Does anyone realise that the theoretical performance of this chip (one) is over 250 gigaflops!!!!?
and there is supposedly 4 (or at least 2?) in a ps3 plus "GPU" (possibly with some traditional parts being taken over by the cell chip)
SO one cell chip == 5* 6800 chips == about 40 or 50 high end pentium/athlon systems in theoretical performance, and even if only half of that can be utilised in real use, thats still 125 gigaflops!!!!!
With four of these chips you have a top 500 supercomputer in a box!!!!

Don't give me all that jibber jabber

[glrotate]

as Mr. T would say.

How about some actual SPECint and SPECfp?

Oh, nothing like that was released? Hmm. makes you wonder. Sort of like the Itanium flop where the excuse, going on for about 10 years now, is that the compiler isn't quite optimised yet.

Any nerd over 15 ought to have heard far to many claims of "revolutionary cpu design" to know better.

Re:Don't give me all that jibber jabber

[tomstdenis]

I think in this case we're talking handwritten asm for the graphics/sound that is re-used in a higher level lang.

Almost like... the way PS2 programs are developed ;-)

Though yeah, lack of oooe [out of order execution] means that as a general purpose CPU the thing will suck bad [because as you translate from HL to opcodes [risc in this case] there are often many cases of parallelisms and renames you can exploit].

Tom

bitter are we?

[I judge you]

Sony claimed that the EE would be twice as fast as a 733Mhz P3. It is at least twice as fast as a 733Mhz P3.

Except that it's by any reasonably useful benchmark. And it's not for any games.

Re:bitter are we?

[be-fan]

That first sentence made no sense, and the second was pointless because gaming performance is a function of the CPU + GPU, so you can't use it to measure CPU performance alone.

Re:Where the fuck are you getting this?

[diablomonic]

actually according to the reports ive read, the cell is supposed to debut in servers long before ps3 arrives, and i remember marketing estimates of 16 terraflops for a 1 rack server!!!!!!

8 cores clocked at 4.7ghz.. So yeah.. fast!

[Viewsonic]

Nuff said. I dunno how Microsoft and Nintendo will be able to compete...

Re:8 cores clocked at 4.7ghz.. So yeah.. fast!

[conna01]

Microsoft will make back the money on the games... From what I have seen, M$ looses money on each Xbox. Nintendo makes good games because they don't try to recreate the "real" world. They make cartoon style games play good and back it up with a story. Brand loyalty will play out...

This is IBM this time, not just Sony.

[Viewsonic]

Not only that, but quite a few different companies in addition as well. So no, I dont think this is some goofyhype machine. IBM doesn't act that way.

You will see this on everything soon enough.

[Viewsonic]

Apple will probably be the first true PC to actually start using it, and then Microsoft will need to port their OS to compete. Makers like Dell and such will probably go out of business as an entire industry will be turned completely upside down. It's that revolutionary.

Re:You will see this on everything soon enough.

[rimmon]

And what should stop Dell et al from producing and selling PCs with a cell processor?

Re:You will see this on everything soon enough.

[nagora]

And what should stop Dell et al from producing and selling PCs with a cell processor?

There is a danger for the tier-1 suppliers that Microsoft drags its feet getting into a new architecture which goes on to beat intel (unlikely to happen, but that might be the very thought that holds MS back). While Mac and Linux systems are appearing using the new chip, they'll be held back by their deals with MS from supplying the hardware (remember: a machine without Windows is a machine for piracy!). It's possible that someone else will dash in and take the new market, leaving the old market to whither away. Dell etc would then be in a very weak position once they broke away from MS's stranglehold on their businesses, or alternatively MS finally realses a CELL-Windows.

It could happen but it's more likely that if CELL takes off, MS will have an alpha version of Windows running on it quite quickly and that will keep them in the game for the next three years as they try to get a version that does more than Minesweeper working. That's basically what happened with Intel and x86-64; Intel was so slow that in any free market they would have been overwhelmed by the competition, who were years ahead of them. But the market isn't free and Intel are almost as strong now as they ever were.

TWW

Re:You will see this on everything soon enough.

[rimmon]

I can (and in fact already have) order systems from dell and various other tier1 suppliers without any OS and preloaded with Linux. Dell even offers a whole bunch services to get you switched to Linux. That is mostly true for servers, but if you're a business customer you can order desktops without an OS just fine, too.
So, please tell me again why Dell is in any risk to go belly up just because a new processor is shipped? You still need somebody who assembles your boxes, ships them and provides you with some kind of warranty and support.
Since lots of the large customers of Dell (and any other tier1 supplier) have contracts with their suppliers (and most large corporations move rather slowly) I don't think much will change.
Besides, I'm pretty sure that Michael has no problems selling everybody brand new PCs and Servers with that revolutinary cell thingy inside. :)

Re:You will see this on everything soon enough.

[nagora]

That is mostly true for servers, but if you're a business customer you can order desktops without an OS just fine, too.

Have you done that?

TWW

Re:You will see this on everything soon enough.

[rimmon]

Yeah, sure. Just go to the Dell site and check the section for large business and corporations and start looking for "DOS Installation" as OS. For example all (or most, to lazy to look) of the OptiPlex Systems are available with said option, some of them are only available that way.

Disclaimer: All of this is true for the european union, if you're from some other neck of the woods ymmv but I'm sure if you talk to your br he's going to make you a deal.

Re:by comparison...

[TheGavster]

That which is written on product labels, sounds profound.

This may be crazy

[pugugly]

I keep looking at this architecture, and although I have absolutely no real qualifications to say why, it keeps jumping out at me.

This thing looks like the architecture is designed with a jump to quantum computing in mind. The methods just . . . bring that to mind.

Which doesn't particularly make sense. Quantum computing is massively efficient at certain specialized problems like searches and of course encryption. I suppose a pathfinding engine with that capability would be extraordinary too, but by and large gaming doesn't have a lot that quantum computing would help all that much with. Perhaps someone smarter than I can explain why this is a spectacularly stupid.

Because it *looks* like something that's designed with that capability in mind. A least to me.

Re:I wonder....

[MP2030]

They already wrote that software four plus years ago, it was called ME.

Re:don't get your hopes up just yet

[iirving]

The New Xbox 2 (or Xbox 360?) is using the PowerPC , if fact Microsoft is currently using Apple G5 as the development platform. So they will have experiance on the Power architecture. I seem to remeber them doing some work in with NT on PCC in 98? but it was killed.

Confused

[BobPaul]

CELL is a Multi-Core Architecture
Contains 8 SPUs each containing a 128 entry 128-bit register file and 256KB Local Store

From your article and the origional it seems there are 8 Synergistic Processing Units/Elements and 1 Power Processor core.

How does s Synergistic Processing Unit differ from a CPU core used in most multicore processors. I'm assuming a SPU doesn't do as much as a CPU, but does anyone know more specifics? Are there benifits other than price for using multiple SPUs and a single CPU rather than multiple CPU cores?

Re:Confused

[Ohreally_factor]

Think vector units, like the Altivec, and you'll be going in the right direction.

Re:Confused

[hattig]

From what I've read, each SPU is its own processor which can address the 256KB of integrated memory only, and presumably send data back into the Cell.

The difference is that this processor has a vector unit instead of a normal scalar ALU. Apparently this vector unit isn't as capable as Altivec/VMX, but presumably includes all the important operations.

What the main PowerPC processor in the Cell does is load each SPU with a program to process streaming data. Each SPU runs the code independently of the main processor. The results from one SPU can be streamed directly into another SPU, hence Cell is a streaming processor. Basically you have an 8 stage software pipeline, where each stage can be a quite complex function (e.g., fast fourier transforms, folding@home, seti, heh).

The main PowerPC core also has a VMX unit of its own, this is a more general purpose vector unit that has full access to the system. I assume that the 2MB of SPU memory on each Cell is memory mapped into the PowerPC's memory space for example. The PowerPC core has SMT (like Intel's HyperThreading), but is rumoured to be cut down from the obvious choice of being a single-core POWER5.

The PS3, as it is currently known, will be a quad-processor PowerPC machine, with 32 vector processors which will obviously be used for graphics and audio applications in the most part, possibly security tasks too, and whatever else can be vectorised.

Expect a 2 or 4 Cell Version.

[Anonymous Coward]

The version of Cell announced today contains eight 64-bit floating point processors.

One way they could easily cut the die size and therefore costs down is by releasing 2 or 4 cell versions.

Thermals

[obeythefist]

To run the latest near-4GHz CPU's from intel, you need to mount a whopping great water cooling system to handle the heat.

Now Sony wants to put 3 or 4 x 4.6GHz CPU's into a single small form factor box?

I seem to remember how loudly the Sony fanboys complained about the noise coming from the XBox... this will be amusing. Will it ship with it's own refrigerator?

Re:Thermals

[nagora]

To run the latest near-4GHz CPU's from intel, you need to mount a whopping great water cooling system to handle the heat.

A massive part of which is the antiquated Pentium (ie, 8086) architecture screaming at the force needed to run at those levels. I doubt that the cell will run as hot as a single P4. I don't think it'll be fanless, of course, but it will be much better off thermally.

TWW

Re:Thermals

[10Ghz]

AFAIK, the x86-ISA takes just few percentages of the die-size. Hardly a "massive part". Modern x86-CPU's decode the x86 ISA in to micro-ops that are then handled by the CPU. Hell, G5 also runs hot, how can that be, since it doesn't have "massive amounts" of x86-legacy to carry around?

Re:Thermals

[obeythefist]

Indeedy, the heat generated/dissapated by a CPU is generated by two main factors, simply put.

1) The amount of electricty flowing through it.
2) The size of the CPU

Now, you need a lot of electricity to run a complex CPU and you need more to run it at higher frequencies. The Cell is going to run at speeds higher than 4GHz, and it's die size is greater than that of a P4. So it's going to need whopping great amounts of electricity to make all that magic happen.

CPU's aren't 100% efficient, nothing is, so some of that electricity will bleed off as heat. Since we're dealing with (presumably) more electricty than a P4 (let's not count idle states needing less power because if the magic of the Cell technology happens, we'll use all the processor can handle), the heat is going to be sensational.

Re:I wonder....

[WMD_88]

Don't forget KDE bloatware.

Wait for the bukkake jokes

[Ohreally_factor]

Oh, great. Future comparisons of computing capability will be based on SPUs. "Hey, I got the new 12-SPU chip in my computer. How much SPU is in your computer?"

My computer has so many SPUs that I need to wear goggles and a raincoat just to run a spreadsheet.

In Soviet Russia, your computer SPUs on YOU!!

Sony Marketing Dept. hires Dr. Gene Ray

[Ohreally_factor]

Nature's HarmonicSIMULTANEOUS 4-Day Time Cell

Believers are ignorant of Cell Creation
and will be destroyed by their own word.

All Creation occurs between Opposites.
On Earth between opposite hemispheres,
and for humans, opposites sexes. The 2
opposite sexes equate to 2 separate Cells,
as if dice of femininity and masculinity -
equating Human Life to a Crap-Shoot
chance of exciting lifetime possibilities.

I have demonstrated absolute proof
of "Cellular Creation", through its
attributes of 4 simultaneous 24 hour
days within a single rotation of Earth.

I have $10,000.00 that I will
wager that Cellism transcends
and disproves Theism Creation.

You ignorant dumbass.

Well, cause Intel is a failure

[chriso11]

Well - let's think about what you are saying. Intel owns the PC CPU market. But everytime they try to get anything else, they get their butts spanked. Remember Intel's graphics? Or their peripherals? Yes, they have chipsets, but their hold is very weak at this point.

The reason that Intel is bashed is cause they have owned x86, and rather than take x86 to the next level, they simply created a marketing chip (the P4, in case you weren't aware). They have become a giant turtle. AMD has gone in and totally flanked Intel. They are the ones doing the more creative implementations, by showing what you can do with x86 - taking x86 to the next level even though it is Intel's turf.

That is why AMD gets kudos while Intel gets bashed.

Re:Well, cause Intel is a failure

[Sycraft-fu]

Hey man, if $34 Billion in revenue is getting your butt spanked, then line me up and get the paddle.

As for graphics, those have been a huge success, to the point that nVidia and ATi began copying the idea. Intel's integrated chipsets are a huge hit with business. They keep costs and space down, and high performance grapihcs aren't necessary for office work. The integrated low-end graphics chip is getting to be quite popular.

Networking would be another huge non-processor area that they excell in. If you ask me what kind of NIC I want in a server, Windows, Linux, BSD, whatever, the answer is Intel. Nobody else I know makes cards of the same quality. 3com used to, but not anymore.

Now the x86-64 thing is an interesting one to pick on, because the reverse is true. AMD was being the uninnovative one. They decided that innovation, in this case, was unnecessary and counter productive. They decided to just whack on 64-bit extensions to the x86 architecture, as was done with the 32-bit conversion years ago, and call it good. It offered nothing new in terms of ISA, but that meant backward compatibility.

Intel tried to be radical. EPIC is a neat idea that's been messed with for years and never made practical. You have the compiler do all the work of deciding what runs in parallel, rather than the chip. Makes for helaciously complex assembly, but that's ok, you just need a good compiler, and Intel makes the best.

Well, total non-starter in the desktop market, that's gone to x86-64 and it's not changing. However seems to be working in the high end computation market. We just got in 2 racks of SGI Itanium coputers for one of the research labs. From what I hear, they are badass number crunchers.

Now if you want to talk some major failures, let's have a look at AMD's motherboard situation. When the Athlon came out it was abysmal. AMD couldn't produce a reasonable chipset to support their own processors. It was slow and incomplete, and couldn't deal with basics like AGP 2x. VIA had a full featured chipset, that was full of bugs and couldn't handle hardware like the GeForce in many configurations. ACPI problems plauged all boards.

Now the point here isn't to try and say Intel's better than AMD. The point is, both companies have hits and misses. Some products can be both a hit in one way, and a miss in another. However there's a lot of fanboyism about AMD and hate towards Intel and its not productive.

You should pick your platform based off of informed choices about what performs better for you, and gives you that performance at the best price. If you find yourself having to justify it by attacking the other company, you probably made it for the wrong reasons.

This goes extra for doublespeak like hating on Intel for focusing on MHz, then hating on them again when someone else does so.

Re:Well, cause Intel is a failure

[warrior]

Intel's Montecito is the best CPU man has made to date. 1.7+ billion transistors, two multithreaded cores, 24MB L3 + 2MB L2I + .5MBL2 cache, 2-something GHz, and it's power efficency is incredible -- something like 100W max power. If the lazy arse programmers would stop suckling up to x86 and adopt a modern architecture we might be able to dump x86 and finally move forward. Fat chance if mainstream will actually adopt a stream processor like Cell. It's a great idea, came out of some research at Stanford. Thing is, running anything other than multimedia is a bitch. If you think compiling code for IA64 is tough don't even bother with a stream architecture. Slashdot continues to suck...

Re:Well, cause Intel is a failure

[TheSunborn]

Best depend on what you need it for. Yes it might be the fastest cpu ever designed, but it is also the most expensive. At a price of the cpu that is atleast 6000$ it is far to expensive. twice the performence at ten times the price makes it diffucult to call it the best.

I actuelly like the Pentium M better. A chip as fast as the 3.0Ghz Pentium IV, but only using 21W and available for 500$ is what I would call a better chip.

Cell's PowerPC core is in-order not out-of-order

[Ideaphile]

I was at the Cell event today, and quoted in some of the news stories. I also have the ISSCC technical papers.

The PowerPC core in the Cell prototype chip is NOT a Power5, as speculated here. According to IBM, this core was designed from scratch for this application. One critical difference is that the new pipeline executes instructions in strict program order rather than reordering instructions to improve throughput as is done with Power5.

Also, IBM has not described the core as "simultaneous multithreaded", just "multithreaded." I presume from this that the multithreading is coarse-grained-- only one thread is active at a time, unlike Power5 which can execute instructions from two different threads in the same cycle.

The logic design for the Cell CPU was optimized for higher clock speeds in a given process than Power5 can achieve. This is a good tradeoff for more linear multimedia algorithms, but reduces effective throughput on other types of code.

I think it's reasonable to suppose that if Apple were interested in using the Cell architecture, it would prefer to use a version of the design that includes a Power5 core in place of the one in the Cell prototype.

. png

mod it.

[cgenman]

Cut an 80x80 hole in the top plastic about the middle of the unit, facing upwards. Mount a lo-flow panaflo over the hole, facing up. Wire the panaflo to an external 9v wall wart. Cut the wire to the original fan. For an added bonus, undervolt the panaflo to 6 volts, and add an undervolted rotary fan into the HDD drive bay. This should be silent overall and have a great airflow.

The default internal fan is a 6v 50mm screamer, narrow and loud. Ultimately it moves less air than any undervolted panaflo. Plus it is pretty terribly positioned. A fan on top of your case may prevent you from stacking other equipment on top, but it is perfect for sucking heat off of the gargantuan main heat sink, and gives the fan enough space that the blades don't make that choppy noise. Thanks to a ludicrous heat sink the PS2 doesn't really require that much cooling, though it does require some. It is usually about 20 minutes before a PS2 with no fan will overheat, while a mostly symbolic fan is usually enough.

Of course, the PS2's are engineered to work in hot climates, so if you're near the sahara you may want to run your fan faster. But overall the systems don't need to be noisy.

*not* Ideal Linux chip.

[tetromino]

There are some problems with using the Cell as a Linux chip. First and foremost, the Cell is a much "dumber" processor than anything that has been on the market for the past 10 years - for instance, it uses the OS (or the application) to manage each core's cache. (Most modern CPU's try to hide the complexity of their internal implementation from software - the Cell, however, requires programs to explicitly use all of its hardware internals) Modifying the Linux kernel so that it can manage nine cache banks while maintaining both good security and reasonable performance will be an interesting research project - and it won't be done in one day.

Add to it the fact that Linux is really designed as a general-purpose OS for general-purpose hardware. Implementing NUMA took years of work from SGI and IBM. I imagine that adding efficient support for the Cell's bizarre asymmetric 9-core design (keeping each core fed, not overloading the Cell's internal bus, etc.) will be a multi-year chore. Then, all the applications (or at least all the major libraries) will have to be rewritten to take advantage of the Cell's design. And even then, gcc won't produce efficient code for it -- you really need a good compiler to take advantage of the Cell, and gcc's main focus is portability, not performance...

I can see something like a PS3 using the Cell -- after all, all the games are hand-coded for the processor, and there isn't much of an OS. I can see IBM using the cell as addon accelerator cards for its AIX systems, since it controls the entire AIX stack. I don't see Linux being able to take good advantage of a Cell system for a while.

"Nazgul"?

[Dr. Cody]

We'll see. IBM has historically been very smart (and sometimes downright ruthless) when it comes intellectual property issues. Their IP attorneys aren't referred to as "The Nazgul" for nothing...

I'm afraid I don't follow--are you saying that they come from R'lyeh?

http://en.wikipedia.org/wiki/Nazgul

[Anonymous Coward]

http://en.wikipedia.org/wiki/Nazgul

P4 double clocked ALU...

[burnttoy]

Large chunks of the integer ALU of the P4 operate at double the clock frequency which is up to 7.2Ghz.... so Cell is "slower" (no not really).

Re:ANY ___FACTS___ AT ALL??

[gl4ss]

they bought into that they could actually use the 6.2gflops meaningfully.

that's exactly what happening now also, journalists trying to put inflated numbers into context in a fashion "what would they do vs. desktop" and they come up on top, while in reality they most likely won't. the hype is exact clone of the hype before ps2 - some journalists never learn(or didn't actually even try ps2 for a desktop. it does shittily as a desktop.).

and on top of it it's going to be hard to program for, not magically easy like blachford makes it out to be - and it's all let up to the programmer!

my point.. just wait for the thing to be REALITY and benchmarkable(as in real world performance, with real production) before salivating over how it will change the face of computing(and at the same time be record cheap somehow despite huge die size).

The death of 3d cards ?

[S3D]

IT may be alittle too far-fetched but the cell processor or it's successor can be a real GPU-killer. May be we will live without 3D videocard in the near future and return of the software rendering. The GPU have performance more then two order of magnitude of contemporary CPU, all due to highly vectorized and paralellized architecture. But if you look at the cell processor architecture you will see it's very similar to GPU. What cell is in fact is some kinde of generalization of GPU architectuire so it could run any kinde of code, not only "per-pixel/vertex", with both read and write dependancy (modern GPU can run only read-dependant shader code). More than that, cell architecture is very well suited for voxel-based rendering. So, not only we could possible be get rid of 3d cards, but voxel rendereds could outperform polygon-based rendereds again (he, he, good buy DirectX)

Re:The death of 3d cards ?

[KirkH]

If that's so, then why is Sony putting an nVidia GPU into the PS3?

I think that using a CELL in place of the GPU may have been their original plan, but it didn't pan out for one reason or another and they fell back on nVidia for Plan B.

Fried eggs, please!

[BigYawn]

Will we be able to use the cell processor as a frying pan, like with the P4?

Re:ANY ___FACTS___ AT ALL??

[be-fan]

they bought into that they could actually use the 6.2gflops meaningfully.

The EE wasn't some magic architecture that could only do 6.2 gigaflops on a single benchmark. It was a simple vector architecture. Developers knew that. If they thought they could get 6.2 gigaflops out of it, but they didn't know how to program vector architectures, then whose fault is that? I don't see anybody complaining about the benchmarks published for things like the Earth Simulator, because it's pretty hard to get close to peak on that thing too!

With regards to what Blachford said --- I refuse to attribute journalistic hype to IBM and Sony. It's the journalists making far out claims, not IBM. A lot of people are legitimately excited about the thing, not because they believe they can actually get 256gigaflops running Quake III, but because they see the 256gigaflops peak, and salivate at the real numbers they could post in their apps.

Re:Clock Speed isn't Everything

[10Ghz]

Well, we already have a good idea how PowerPC performs (although the PPC in the Cell might not run as fast as rest of the chip does). As to the SPE's... Ars Techinca says this:

The SPE's very simple front end can take in two instructions at a time, check to see if they can operate in parallel, and then issue them either in parallel or in program order. These two instructions then travel down one of two pipes, "even" or "odd," to be executed. After execution, they're put back in sequence (if necessary) by the very simple commit unit and their results are written back to local memory. The individual SPUs can throw a lot overboard, because they rely on a regular, general-purpose POWERPC processor core to do all the normal kinds of computation that it takes to run regular code. The Cell system features eight of these SPUs all hanging off a central bus, with one 64-bit POWERPC core handling all of the regular computational chores. Thus all of the Cell 's "smarts" can reside either on the PPC core, while the SPUs just do the work that's assigned to them....

Once the instructions are in the SPE, the SPE's control unit can issue up to two instructions per cycle, in-order. The SPE has a 128-entry register file (128-bytes per entry) that stores both floating-point and integer vectors. As stated above, there are no rename registers. All loop unrolling is done by the programmer/compiler using this very large register file.

Re:Oh sweet!

[Zoolander]

Obligatory Gentoo compile time joke... check.

OSS to change the playing field?

[jago25_98]

Why not completely design the lot, run OpenSource on it because that's more portable, market to servers as they run samey applications,

and then go on from that.

I guess no one wants to take on x86; i.e. ARM.

Re:DumbassThisIsIBMUfool

[glrotate]

Would a better comparision would be the PPC 615?

Imagine...

[Torontoman]

Imagine a beowolf cluster of these! Legit comment actually.

So try offloading it to software

[Short Circuit]

No oooe? Build a really, really good compiler.

Have an app originally compiled inefficiently? Run code-reorganizing tool on it.

Not sure that any of the latter tools exist today, but it's something I've been pondering. Imagine, on a PC, replacing a bunch of SISD instructions with the latest SIMD equivalent.

The hard meat and veg in more ways than one

[tod_miller]

Imagine a beowulf cluster of these!! no really! FUCK ME!!!!

The first version of the chip will run at speeds faster than 4GHz. Engineers were vague about how much faster, but reports from design partners say 4.6GHz is likely. By comparison, the fastest current Pentium PC processor tops out at 3.8GHz.

Cell can process 256 billion calculations per second (256 gigaflops), falling a wee bit short of marketing hyperbole calling it a "supercomputer on a chip." The slowest machine on the current list of the Top 500 supercomputers can do 851 gigaflops.

The chip will have 2.5MB of on-chip memory and have the ability to shuttle data to and from off-chip memory at speeds up to 100 gigabytes per second, using XDR and FlexIO interface technology licensed from Rambus.

"One of the key messages you hear from the architects of next-generation chips is that their performance is being limited by off-chip bandwidth," said Rich Warmke, product marketing manager at Rambus. "We've really licked that with Cell. One hundred gigabytes per second is really unprecedented in the industry."

The chip will have 234 million transistors, measure 221mm square and be produced using advanced 90-nanometer chipmaking processes. Peter Glaskowsky, a consultant at research company The Envisioneering Group, said he expects Cell production to shift to a 65-nanometer process, however, as IBM introduces the chipmaking technology later this year.

I have so got a boner right now. I want 2 of these.

Cell is not an x86 competitor.

[flaming-opus]

It is a radeon/geforce competitor. Or something like that.

The cell processor is only really fast when the spus are in use, which means 32-bit non-branching floating-point arithmatic. For anything involving integer math, flow control, or uneven memory access, the SPUs defer to the main processor. I'm sure IBM put a decent processor in there, but it doesn't sound like it's anything revolutionary, and there's only the one.

What does this get you? -- A processor that is really good at decoding mpeg, rendering graphics, maybe approximating the physics of flying dragons. It is not a fast general purpose processor. Operating systems, word processors, databases, these are all integer tasks, and much more-so they are branch tasks. Scientific computation - this requires double-precision floating point. Photoshop is about the only piece of non-multimedia software that might be able to take advantage of this.

The end result is that this will likely be a great chip for set-top boxes of all sorts, maybe even for video-editing workstations. A G5/pentium replacement it isn't; that's a different ball game.

Re:Cell is not an x86 competitor.

[be-fan]

The cell processor is only really fast when the spus are in use, which means 32-bit non-branching floating-point arithmatic.

Actually, the APUs handle both floating-point and integer vector arithmetic (like Altivec or SSE2). And it has a branch unit, and since its so simple, it's pipeline is probably a reasonable length. So it would likely be good on branchy integer code, as long as it stayed within a small dataset. I can think of uses for that --- certain fluid dynamics, structural analysis, or thermal analysis problems could fit nicely. Also, it could be good at solving constraint systems, numeric integration, and a number of other algorithms where runtime can be quite large for small datasets.

PS3 "comparative power" over its lifetime

[Savage650]

PS3 [..] may have a lifespan of 5-10 years. So you have to imagine that in 2017 [..] These consoles have to be armed with the best possible components in order to give them longevity and flexibility in the market to compete with Xbcube units produced by Nintendosoft in 2010. ;)

Good point. But Sony has another ace up their sleeve: while the PS3 might or might not offer internal CPU upgrades, the Cell programming model is meant to scale well for networks/clusters of Cell-enabled devices. Thus, you should be able to harness the cpu power of other Cell-devices in your home (another PS3, a settop box or even the new Sony TV).

I'd expect the "Flight Simulator-with-real-cockpit" crowd to go for a solution based on multiple PS3: every box handles part of the IO (tons of buttons&blinkenlights) and some 2D graphics (instrument panels, MFDs). The remaining CPU capacity is pooled towards rendering the outside view(s).

Re:Cell's PowerPC core is in-order not out-of-orde

[EvilTwinSkippy]

The PowerPC instructions core is simply to provide a way to leverage existing compilers against this new architecture. Sony learned, the hard way, that developers don't want to hear about a quirky new instruction set.

I hear you though. The Power5 is designed to handle large multi-process loads. This new Cell architecture, or at least this particular Cell chip, is designed for real time processing of large piles of data.

I'm not reliving computer architecture class... I'm not reliving computer architecture class... (open's eyes) ... Whew

Gigahertz Schmigahertz

[PeanutGallery]

From the Intel days, the press and the populace has been hypnotized into thinking "processing power=Ghz. End of story". The fact is if the cell specs are more than just hype:
1) 64 bit architecture
2) 8 of those on a chip
3) you can gang Cells together via boradband

A cell running just 1Ghz would easily blow the pantaloons off anything we've got in our house today!

Watch for Cell-powered Macintosh next year.

[MtViewGuy]

I have this feeling that Apple will announce that they will switch to the Cell CPU some time later this year, and the first machines to use the Cell CPU will be announced at MacWorld in January 2006 (probably both Power Macintosh and iMac models).

Re:Idiot

[Sebastopol]

my CPU is making noise right now, it's saying, "anonymous coward loser alert!"

Money quote from Ars conserning Cell Development

[CosmicDreams]

source: http://arstechnica.com/articles/paedia/cpu/cell-1. ars/2

To sum up, IBM has sort of reapplied the RISC approach of throwing control logic overboard in exchange for a wider execution core and a larger storage area that's situated closer to the execution core. The difference is that instead of the compiler taking up the slack (as in RISC), a combination of the compiler, the programmer, some very smart scheduling software, and a general-purpose CPU do the kind of scheduling and resource allocation work that the control logic used to do.

sorry if someone else has posted this already.

Re:I wonder....

[bbrack]

Cell would probably not make the best server processor. Although its performance has been tweaked to be able to vector calculations EXTREMELY quickly, it looks like the only thing it has in its favor as a server processor is the extremely high memory BW. A better example of a server processor designed under the same type of methodology (lots of little parts doing things in parallel) would probably be Sun's Niagara - similar design concept to cell, but teaked to perform server operations, not floating point vector calculations, very quickly.