SPEs and PPEs in cell

Are they some sort of cores or are they some sort os co-processors,

Thanx

Cell is 9 core CPU, but all the cores are not the same and have different functionality, the inventors are thinking about the asymetric multicore architecture that is the future and other compainies like Intel and AMD will follow their example.
The PPE, named as Power Element is 1 core that is very similar to PowerMac G4 processor. It has the role of accepeting the tasks and share them to the rest 8 cores, named as Synerghistic Cores.
The PPE is properly sharing the multimedia jobs to the SPEs, taking in account which one is how available. It is easy to divide and process multimedia in parallel, but the PPE task is not that easy becouse there are a lot of data that is flowing through it.
SPE 8 cores are connected with extremly fast links with bandwidth of 300GB/S per link, 256kB cache each, operating at 4-5GHz, therefore doing 192-256GFLoop/s. That is...OMG, my overclocked Athlon is doing less than 6GFloop/s

IEEE Spectrum, January 2006
MICROPROCESSORS By Samuel K. Moore

"WINNER MULTIMEDIA MONSTER: Cell's nine processors make it a supercomputer on a chip"

GOAL: Make a new microprocessor architecture that beats all others at handling graphics and broadband multimedia.
WHY IT'S A WINNER: It met that goal and is being designed into highvolume massmarket items like game consoles and televisions.
ORGANIZATIONS: IBM, Sony and Toshiba.
CENTER OF ACTIVITY: Austin, Texas.
NUMBER OF PEOPLE ON THE PROJECT: 400 at its peak.
BUDGET: US $400 million.



WE'RE FLYING AT ABOUT MARCH 1.5 around Mount Saint Helens, in Washington state. IBM Corp. senior programmer Barry L. Minor is at the controls, rocketing us over the crater and then down to the lake at its base to skim over the tree trunks that have been floating there since the volcano exploded over 25 years ago. The flight is exhilarating, even though it's just a simulation projected on a widescreen monitor in a cluttered testing lab.

Then, at the flick of a switch, Minor turns the simulation over from his new Cell processor to a dualprocessor Apple Power Mac G5, and the scenery freezes. The G5 almost audibly groans under the burden, though it's no slouch. In fact, it's currently the top of the line for PCs. But Cell is something different entirely. It's a bet on what consumers will do with data and how best to suit microprocessors to the task and it's really, really fast. Cell, which is shorthand for Cell Broadband Engine Architecture, is a US $400 million joint effort of IBM, Sony, and Toshiba. It was originally conceived as the microprocessor to power Sony's third-generation game console, PlayStation 3, to be released this spring, but it is expected to find a home in lots of other broad-band-connected consumer items and in servers too.
Executives at Sony Corp., in Tokyo, wanted more than just an incremental improvement over PlayStation 2's processor, the Emotion Engine. What they got was a 36fold acceleration, to a whopping 192 billion floating-point operations per second (192 gigaflops). Because Cell is a combination of general-purpose and multimedia processors, it defies an exact comparison with other upcoming chips, but it's thought to be more powerful than the chips driving competing game systems.
Cell can calculate at such blazing speed, in part, because it's made up of nine processors on a single chip of silicon, optimized for the kind of realtime calculations needed in today's broadband, media-rich environment. A specially designed 300gigabit-per-second bus knits the processors into a single machine, and interface technology from Rambus Inc., Los Altos, Calif., gives it fast access to memory and other offchip systems.
So far, microprocessor watchers have been impressed with what they've seen of Cell. "To bring huge parallel processing onto a single chip in a clean and efficient way is a real accomplishment," says Ruby B. Lee, a professor of electrical engineering at Princeton University and an IEEE Fellow.
A graphics-heavy item such as PlayStation 3 isn't just a showcase for an unusual chip. For IBM it's a philosophical statement. "Gaming is the next interface driving computing," says James A. Kahle, Cell's chief architect with the IBM Technology Group, in Austin, Texas [see photo, "Multicellular"]. Just as moving from punch cards to electronic displays changed what people expected of computers, the highly collaborative, realtime realism of today's games will set the standard for what people want from computers in the future.



But even now, the sheer desire for power in the gaming market guarantees that Cell will he made in volumes that more than make up for the loss last year of IBM's highest profile customer, Apple Computer Inc. Market research firm iSuppli Corp., in El Segundo, Calif., predicts that 37 million game consoles will be sold this year alone worldwide. By 2007, when all three game console makers will have released their nextgeneration products, the market will have grown to 44 million. And though Cell is exclusive to the PlayStation 3, IBM has a lock on the rest of the console market. Its microprocessors will power both of Sony's competitors, Microsoft's Xbox and Nintendo's GameCube.
The Cell powered PlayStation 3 can expect to pick up a little less than half of what could become a market worth up to $9.5 billion in 2007, according to iSuppli senior analyst Chris Crotty. And, of course, there are other high-volume plans for Cell.
Toshiba Corp., in Tokyo, for one, plans to build television sets around it. The company has already shown that a single Cell processor can decode and display 48 compressed video streams at once, potentially allowing a television viewer to choose a channel based on dozens of thumbnail videos displayed simultaneously on the screen. And in a smaller market, Cell has already found its first outside customer in medical and militarysystems maker Mercury Computer Systems Inc., in Chelmsford, Mass., which is developing a two Cell blade server due out by April.
With two such massive consumer electronics makers as Toshiba and Sony behind it, Cell is an obvious attempt to control the "digital living room," as technology executives have dubbed their dream of a home where all the media players are intelligent and networked together. "[Sony's] goal is to make a computer fun...to make it an entertainment platform," says Sony's Cell director Masakazu Suzuoki. "But even if we make the Cell system an entertainment platform, there's nothing if there's no content."
Indeed, experts say Cell's success hinges on whether programmers outside IBM, Sony, and Toshiba will be able to exploit the gigaflops that Cell has to offer. Tony Massimini, chief of technology at the consulting firm Semico Research Corp., in Phoenix, puts it bluntly: "Cell has strong potential, assuming that the game developers satisfy their customers' needs. But if the games suck, who wants to buy it?"
That Cell has more than one processor core on a single chip is more a sign of the times than a revolution. All the microprocessor stalwarts are moving to multicore design. The principal reason is that the old way of doing things—increasing the number of calculations per second by shrinking the processors into a tighter knot of tinier transistors and then dialing up the clock speed has essentially crashed headlong into the brick wall of heat generation.
Because transistors using today's technology are so small, even when they are supposed to be in the "off" state, infinitesimal currents still leak through them. That leakage warms them constantly, and with the extra heat generated when transistors switch "on" or "off," it produces a microfurnace on a chip. If chip makers had continued on their old path, by the year 2015, microprocessors would be throwing off more watts per square millimeter than the surface of the sun.
As a result, the industry has shifted from maximizing performance to maximizing performance per watt, mainly by putting more than one microprocessor on a single chip and running them all well below their top speed. Because the transistors are switching less frequently, the processors generate less heat. And because there are at least two hot spots on each chip, the heat is spread more evenly over it, so it's less damaging to the circuitry and easier to get rid of with fans and heat sinks.
Multicore processors on the market today are generally symmetrical—that is, they have two copies of essentially the same core on one chip. Cell, on the other hand, has an asymmetric architecture that contains two different kinds of cores [see photo, "Cell City Map"]. One, the Power processing element, is similar to the CPU in a Mac, it runs the Linux operating system and divides up work for the other eight processors to do. Those eight—called Synergistic processing elements—are designed specifically to juggle multimedia applications: video compression and decompression, encryption and decryption of copyrighted content, and, especially, rendering and modifying graphics.



The Synergistic elements were built from the ground up to do what are called singleprecision floatingpoint calculations—the kind of operations needed for dazzling threedimensional graphics and a host of other multimedia tasks. The design traded flexibility-a Synergistic element is not versatile enough to run the Linux operating system on its own—for eye-popping speed. When pushed to its 5.6gigahertz limits, a single unit can do 44.8 billion singleprecision floatingpoint calculations per second. Not wanting to cut Cell off from a role in scientific computing, its designers included circuitry in each Synergistic element that can do the more exacting calculations, called doubleprecision, that scientists demand, but its performance is only about onetenth that of the singleprecision unit.
In fact, the Synergistic elements are so fast that a single one could easily consume the entire bandwidth on the interconnects to the offchip memory, leaving its siblings starved for data and stalled out. IBM and its partners had to design a special chunk of circuitry into Cell just to prevent that problem.
Apart from its raw power, Cell has content-protection tricks that should make it attractive to multimedia applications makers. For instance, the Synergistic element's architecture prevents any application or external device from accessing the element's local memory, so that, for instance, a program cannot steal a music file that is being decrypted by the processor. "Once you bring your code in and decrypt it, it can execute in a virtually trusted environment," says IBM's Cell architect Charles R. Johns. "All the data it calculates on, sends out, and brings in is fully protected."
The isolation function can be used in several ways, says Kahle. "We knew we couldn't anticipate all the different security needs in the future, but we wanted to know we had the right hardware to support a very robust security system."
Barry Minor's Mount Saint Helens simulator is a good example of how Cell's different processors work together. His program takes a satellite photo of the volcano, lines it up with an elevation map, and then turns it into a detailed 3D terrain on the fly. The Mount Saint Helen's data has a resolution of 2.4 meters. The city of Austin, where the Cell design center is, once gave Minor access to its 15.4centimeter-resolution satellite map. "You could land in Michael Dell's backyard and check out his view," Minor says with a grin.
What's happening inside the processor is a finely choreographed dance. The Power processing element starts by figuring out where the joystick is pointing the simulator in the stored 2D maps. Then it divides that scene into 32 portions, four for each Synergistic element. Though perfectly capable of it, the Power processing element does no calculations on the actual data. Instead, it plays to its strength as a controller, figuring out which chunk of work should go to each of the other cores according to how complex the scene is and which cores have more or less time on their hands.
The Synergistic elements then go to work. They pull their portion of the data into their local memories, which they can access at great speed. Then each runs a rendering algorithm on the data and stores it off the chip in the system memory. When the processors are done, they signal the Power element, which instructs one of the synergistic units to run a video compression algorithm. That processor compresses its sister units' finished products and then pushes them out to be displayed on the screen or streamed to a PDA or some other device.
Because the compression takes less time than rendering the graphics, the compressing processor automatically switches gears when it's finished and runs the rendering algorithm on a portion of data until it's needed for compression again. With each frame, the process starts over.
This dance works so well for two reasons. The first has to do with the way Cell handles memory. Rather than waste several clock cycles waiting for the right data to arrive from memory, a Synergistic element works only on data stored in its own 256 kilobytes of memory, to which it has a high-bandwidth connection. More important, Cell's memory-handling engines can be programmed to keep data streaming through the processor. "We can get over 128 memory transactions going in flight at once," boasts Michael N. Day, a distinguished engineer at IBM.
The memory-access engine takes in new data and sends out the old just in time for the synergistic unit to perform the necessary calculations. When Cell runs Minor's volcano simulator, it waits for data to arrive from memory for only 1 percent of the time, the G5, in contrast, stands idle for about 40 percent of the time.
Cell's other key to speed has to do with breaking problems into parts that can be done in parallel. In Minor's simulation, it probably seems obvious that an image can be divided up into eight strips and these worked on independently. What wasn't so obvious was that the 3-D rendering could be done four pieces of data at a time within each synergistic processor. Such four-way parallel computing is called single instruction multiple data, or SIMD, and it is particularly well suited to the manipulation of graphics and other multimedia.
In these problems, you typically want to perform the same operation on each of the elements in a large chunk of data. For example, to increase the brightness of an image, you'd want to add the same number to every pixel in it. Since around the mid-1990s, general-purpose processors such as the Intel x86 architectures have been doing SIMD computing using a set of multimedia-specific instructions, explains Princeton's Lee, a multimedia instructions pioneer.
But SIMD instructions run far faster on Cell's Synergistic processors, because the Cell processors were designed from the start to handle them. And don't forget: there are eight such processors on each chip. Cell programmers spend most of their time turning complex algorithms into efficient SIMD algorithms, says Minor. "Once you've done that, you're 80 percent done."
The Chip's commercial success will depend on whether programmers can learn to exploit its full potential. To that end, the developers have from the beginning put a high priority on crafting the appropriate software tools.
One of the key deadlines the Cell development team had to meet was having its software ready and tested in time for the arrival of the first chips, in spring 2004. The software team was running programs on a Cell simulator two full years before it got the first chip—and when the chip finally arrived, both the operating system and the applications worked on the first try. "Had we waited to do software development until the chip came back, it would have been a disaster," says Theodore R. Maeurer, software manager at IBM.
With such a head start on the software, the group could focus on how to familiarize new programmers with Cell. "A programmer has to do a really nice job of laying out the data transfers and so forth," says Day. But soon that job will be turned over to the compiler and the programming tools. IBM software engineers are also developing tools that will make it easier for programmers to divide tasks between the Power element and the Synergistic cores, and they're making others to automatically find solutions to problems that fit well with the Synergistic units' SIMD strengths. The company has already released more than 700 pages of documents to applications developers and will begin releasing tools and compilers, as well.
Cell's asymmetric architecture signals the beginning of a big shift in how computers are programmed, says Craig Steffen, a senior research scientist at the National Center for Super-computing Applications, Urbana-Champaign, Ill., who gained some fame lashing together 70 PlayStation 2 consoles to form a $50 000 supercomputer.
"How do you program with eight engines running full speed without them constantly stopping and waiting for data?" Steffen asks. Cell will force mainstream programmers to wrestle with that question. But ultimately, parallel programming will become fairly routine, he predicts. "Over the next several years, we won't think of an asymmetric processor as anything different."
Indeed, some think Cell is an indication of what's to come in other microprocessors. "In the future, we'll see convergence of general-purpose multiprocessors and game- and media-oriented processors," says Princeton's Lee. "Media processors will become more general purpose, and general purpose, more multimedia." And with any luck, that will make your living room a more entertaining place.

Your forgot to mention it's difficult to code for and it additionally sucks, max peak performance is attainable on "hand picked code".

We're talking peak theoretical performance; real numbers won't be nearly as high.

-Anandtech.com

It'd be good for BOINC/Folding =D

Yeah, unfortunately it looks like Cell is having many problems on many levels.

Aside from the difficulty to program, it appears that due to the complexity and disperse geometry of functional units on the die surface, actual production yields have been dismal.

Rumor has it that Sony can't even get a single unit to show at E3 this year.

Cell should work well for many workloads, but I remember the initial press blitz when IBM announced the existance of this chip. As cool as it would be, it certainly will never live up to that inital hype.

Maybe the Octopiler will help things a bit.

http://news.zdnet.com/2100-9593_22-6042132.html

Cell will be great for physics but that gimick will get old soon. I heard a quote from a developer saying that you could always tell which room had PS3 development going on because you could hear the programmer's yelling !@#$ and !@#!@$ and Mother!@#!@#$!@#$ sony!!!

Same problem the Itanium has.

Luckily in games it's fairly common to write vectorized code (i.e you can do a 3D dot product in one cycle or a matrix multiplication in 4-6 cycles). The Cell couldn't be any harder to programmer for than the PS2's VU units...

Past vectorizing compilers have failed to optimize code properly (see Itanium compilers and products like VectorC) and probably won't until compiler research makes bigger breakthroughs.

Games, rendering, and high-end scientific apps are all about these "vector" computations. Still, the Cell is cheap and it does double-precision operations fast (single precision floats aren't accurate enough) - it's going to make it in the non-games sphere.

Good for physics ?, how can a completly in-order processor be good for physics, I dont care how many cores there are, physics is one of the most sporratic gaming effects created, you cannot turn them off when the player turns around, you cannot control which way,degree,speed the player chooses to use these physics, and in-order is simply not qualified for that kind of processing environnement. The xbox 360 is ahead of the ps3, I could care less, what cgi trailer sony puts out, they have one real in-game demo and that is I-8 or resisitance fall of man which fails to reach HL2 graphics.... it falls way short of the bar. Since the xbox 360 has the R600 archetecture one may assume it is ready for DX10, if not can be updated with an auto firmware patch via xbox live. Dx 10 is much faster, more optimised. To date no game on the 360 uses all three cores. There is a really good reason why many xbox games ran at 30 fps.... they were simply much more graphically advanced then the ps2. Id really like to see a cell platform in a windows/linux environnement and see the cell get destroyed by our dual cores of today.

It could make a graphical improvement, maybe not higher framerate but the games would look better while maintaining a relativily high framerate. Computer is and will always be better then consoles, for gaming and everything, no matter what sony, ms, or nintendo says. We are getting atis 2nd gen of the R600, which has a rumored 64 half pipes, its going to be a monster. Yea there are alot of xbox games that the framerate would drop off randomly, very frustrating and I didnt know why it did so, good to find an answer. The thing about consoles having higher memory bandwidth is they need it more then pcs, where as they have much lower capacitys of 512 mbs for the whole console (ps3,and 360), we in high end gpus have 512 mbs of memory for just the gpu which is now in the range of 50 gb/s, plus most high end systems have 2 gbs or more ram, which with AM2 coming out,and most intel platforms already having ddr2 memory, most of the memory in a high end pc will be ddr2 and get around 10 gb/s or more memory bandwidth while having 4x-8x the capacity. The new consoles do have alot of cores, but they all are in-order creating an extreme bottleneck.

I will dissagre with you. Very simple, there is no use of any hardware if it is not software supported. It is like you have a ferrari, but you don't have fuel.
Remember what kind of stuff programers did on C64, Amiga, Atari?
What kind of hardware were they using for that?
So, compared to what we have today and todays hardware, I think that software lacks a lot behind hardware.

That's a naive answer.

Consoles have (and still do) a huge bandwidth advantage. While you think this is because of the limited memory, that's not entirely accurate. PCs have the CPU + GPU power, but have traditionally suffered in high-framerate 2D or racing games that require constant streaming of data.

Also Windows is not such a great real-time OS (IRIX is) and pre-emptive multithreading wreaks havoc (creates input latency, frame drops). So really, the modern PC architecture is fine - but Windows is a so-so operating system for games.

Consoles are also standardization and "easier" to optimize for. Console CPUs typically only accept CACHE ALIGNED data. They halt execution if your data isn't on aligned boundaries. They might be slower per clock than PCs CPUs, but the code is always written "correctly."

And until recently, it was impossible to edit/composite film resolution images in real-time on the PC due to bus bandwidth limitations - whereas SGI was doing this back in 1998 (albeit you had to pay a million bucks for this right). SGI used Cray's insanely fast "crossbar" architecture after(?) they bought that company in the mid-90s. AMD's using a variation of it nowadays.

Most console games have a lower framefrate because they are locked at a multiple of the TV's vsync (60hz or 50hz). You get shearing if you display graphics at a different framerate.

I recall the PS2 had approx an average 9 GB/s(?) of bandwidth. It was insane compared to the shoddy PCI bandwidth(~110MB/s) of yesteryear. Xbox 1 had HyperTransport to the I/O system and something else to the GPU.

PCIe x16 is about 4Gbp/s (~ 512MB/s) + FSB memory 3.2-6.4GB/s - all of which combined is mediocre compared to the SGIs.

From wiki:

Xbox: Theoretical Memory Bandwidth: 6.4 GB/s.
PS2: Memory Bus Bandwidth: 3.2 GB/s + DRAM Bus bandwidth: 47.0GB/s.

Fast low-latency RAM, fast caches, etc. all help.

As for modern consoles, I dug out a few numbers:

http://xbox360.ign.com/articles/617/617951p3.html

The PS3 has 22.4 GB/s of GDDR3 bandwidth and 25.6 GB/s of RDRAM bandwidth for a total system bandwidth of 48 GB/s.

The Xbox 360 has 22.4 GB/s of GDDR3 bandwidth and a 256 GB/s of EDRAM bandwidth for a total of 278.4 GB/s total system bandwidth.


The GameCube had super-fast video RAM and losts of bandwidth. It suffered due to the tiny amount of video RAM tho'.

One CPU. One GPU. One type of optical drive. It's MUCH easier to optimize and tune for a stable, predictable platform than the mishmash of PC components - most PC games aim for a low common denominator.

Maybe true, but I have to disagree on what regards past references and near-future probabilistic trends: From a few ground-breaking non-x86 ISA uArchs (Itanium, MIPS, Sparc, Transmeta, POWER, PPC,...), only DEC Alpha vanished (not to oblivion, though: It's still a reference!). Not because the lack of (formidable) compilers, not because of lack of comparable performance, not because of software support (and, even x86 emulation, "FX32!", if I recall correctly...); rather, because it couldn't compete - at the time - with the x86 ISA (well, one of the reasons, anyway).
MIPS adapted & became a hardware-wise open source processing platform; Itanium survives and it's not a soon-to-be paper release; it's a [material] fact.
Cell may be hard to programme; well, x86 is not easy, at all! Although inherently hard, it's stands out in complexity by comparison with x86 & PPC, for example.
As for its workings, it's also advisable to check the sources: http://www-128.ibm.com/developerwo [...] larnutter/

And, it has already left a "material" footprint (aside the PS3):
http://www.research.ibm.com/cell/w [...] _cloth.pdf;
http://techon.nikkeibp.co.jp/engli [...] 25/105050/

Certainly it might vanish, leaving room for more realistic (read actual) approaches (I don't pretend to sound like Cell's attorney!); however, most probably, highly complex compilers will have to be developed, for the near-future uArchs; and Cell might well be one of many new computing paths to explore.
This is my opinion & my conviction.


Cheers!

So, do you agree that software is laging?

So, do you agree that software is laging?

Absolutely.
My point is, because it's lagging (an lacking!) and, because I tend to believe corporations don't just throw away thousands of millions (in money) and all sorts of resources just for the fun of it, the end result challenge is big enough for programming developers to engage into the "do the impossible" task.
Moreover, when we're talking about Microsoft, Toshiba, IBM, Sony & the likes, I risk to affirm that they know what they're doing... and, market-wise, both XBox & PS3 consoles may be a big plus for Microsoft & Sony; all the other intervenients, await success in not-so-specific niches.

I must also state that, not being an expert in any hard/soft areas, I really do appreciate your inputs.


Cheers!

I think I was comparing PS2 CPU+GPU bandwidth(well over 6.4GB/s) to PCI + GPU bandwidth(total bandwidth under 6.4GB/s). circa 2002. The PS2 owned back then because it had to - there was so little memory in the subprocessors: GS, VUs - that the machine was constant sending gigabytes of data around. Even the lowly Xbox had 6.4GB/s in it's unified memory architecture though in practice I think I only got around 4-5 GB/s.

The PS3 and Xbox360 raise the bar even higher. The bandwidth to handle rendering a 60fps 720p 4xFSAA with a million triangles and textures with 512mb shared memory is insane. PCs tend to "cache" vertex and texture data in the GPU reducing the need for a crazy CPU->GPU (PCI) bandwidth consoles, but that means preloading data whereas many consoles have optied for dynamic streaming model (severly limited by the DVD read speed).

Still mediocre compared to a 5 year old SGI server. Opterons are sweet and the fact they are affordable is great, but consoles use the same architecture as the SGIs did - Unified Memory Architecture (no need for separate video and CPU memory), Crossbar/HyperTransport and fast, fast GPUs + rasterizers.

The whole idea PCs have always had greater bandwidth is false. I wanted to show PCs in the past have had shoddy bandwidth and that the new PCs (Opterons w/ multiple HyperTransport links) and SLI/CrossFire have finally caught up to consoles + SGIs.

I hope to see the next-gen of PCs go Unified Memory Architecture - I never liked the idea of separate memory for audio, video, CPU. I also want Opteron-like HyperTransport (I own a dual Opteron rig) bus between all the components: CPU, GPU, Physics processor, audio processor, etc..

AGEIA claims that the PhysX card has nearly 2TB/s of internal bandwidth.