The Mystery of Cell Processors

LucidBeast writes "Consumer appliances requiring more computing power Sony, IBM and Toshiba started 2001 developing "Cell"-processor that comprises of multiple processor cores and should give performance ten times of conventional processors. Now the CNN Money reports that details of the processor will be released Feb. 6-10 at the International Solid State Circuits Conference in San Francisco. Also reported by EE Times. Rumors also tell that Sonys PS3 development platform has already been shipped to some developers equipped with the cell processor."

Comments

Article text

[mrhandstand]

Chip power, times 10
Sony, IBM, Toshiba disclose details of new processor that will run next-generation electronics.
November 29, 2004: 6:13 AM EST

TOKYO (Reuters) - IBM, Sony Corp. and Toshiba Corp. on Monday unveiled some key details on the powerful new "Cell" processor the three are jointly producing to run next-generation computers, game consoles and TVs.

Cloaked in secrecy and the object of much speculation since the three conglomerates announced the project in 2001, Cell will be 10 times more powerful than conventional chips and able to shepherd large chunks of data over broadband networks.

In a joint release, the three firms gave a glimpse of their respective plans for Cell-powered products, but were mum on technical details, which will be revealed Feb. 6-10 at the International Solid State Circuits Conference in San Francisco.

IBM (Research), Sony (Research) and Toshiba are investing billions of dollars to develop and prepare for mass production of Cell, which is a multicore semiconductor composed of several processors that work together to handle multiple tasks at the same time.

"In the future, all forms of digital content will be converged and fused onto the broadband network," Ken Kutaragi, executive deputy president and COO of Sony, said in the release. "Current PC architecture is nearing its limits."

IBM said it would start pilot production of the microprocessor at its plant in East Fishkill, N.Y., in the first half of 2005. It will use advanced 300 millimeter silicon wafers, which yield more chips per wafer than the 200 mm kind.

It also announced plans to first use the chip in a workstation it is developing with Sony, targeting the digital content and entertainment industries.

Sony said it would launch home servers and high-definition televisions powered by Cell in 2006, and reiterated plans to use the microchip to power the next-generation PlayStation game console, a working version of which will be unveiled in May.

Toshiba said it planned to launch a high-definition TV using Cell in 2006. Top of page

Re:Please Help!

[oexeo]

It's been granted moderation powers as well, so be careful what you say

A bit more on PS3

[Sai Babu]

But UNC's Zimmons has his doubts. "I believe that while theoretically having a large number of transistors enables teraflops-class performance, the PS3 [Playstation 3] will not be able to deliver this kind of power to the consumer," quoted from /. referenced article.

Zimmons talks the details.

Re:Future

[Anonymous Coward]

You mean, like a MOS 6502 processor? I've never checked it, but I've been told that's what the code that was shown in the "terminator view" in Terminator 1 was: 6502 assembly. Did they update it in any of the later movies? Say, to a 6510 or a 68510 ?

past.

[leuk_he]

or take a look at the 2002 patent application:

http://theinquirer.net/?article=19941

Re:past.

[Leonig Mig]

http://appft1.uspto.gov/netacgi/nph-Parser?Sect1=P TO2&Sect2=HITOFF&p=1&u=/netahtml/PTO/search-bool.h tml&r=1&f=G&l=50&co1=AND&d=PG01&s1=20020138637&OS= 20020138637&RS=20020138637

from the US Patents Site.

Re:past.

[WillWare]

That's an article mentioning the patent application. The application itself (not including diagrams) is here.

"Cell Processor Unveiled" form physOrg

[S3D]

Cell Processor-Based Workstation Prototype
The companies expect that a one rack Cell processor-based workstation will reach a performance of 16 teraflops or trillions of floating point calculations per second.
Cell Processor Unveiled
IBM, Sony Corporation, and Toshiba Corporation today unveiled for the first time some of the key concepts of the highly-anticipated advanced microprocessor, code-named Cell, they are jointly developing for next-generation computing applications, as well as digital consumer electronics.
Specifically, the companies confirmed that Cell is a multicore chip comprising a 64-bit Power processor core and multiple synergistic processor cores capable of massive floating point processing. Cell is optimized for compute-intensive workloads and broadband rich media applications, including computer entertainment, movies and other forms of digital content.
Other highlights of the Cell processor design include: -- Multi-thread, multicore architecture. -- Supports multiple operating systems at the same time. -- Substantial bus bandwidth to/from main memory, as well as companion chips. -- Flexible on-chip I/O (input/output) interface. -- Real-time resource management system for real-time applications. -- On-chip hardware in support of security system for intellectual property protection. -- Implemented in 90 nanometer (nm) silicon-on-insulator (SOI) technology. Additionally, Cell uses custom circuit design to increase overall performance, while supporting precise processor clock control to enable power savings.
IBM, Sony Group and Toshiba will disclose more details about Cell in four technical papers scheduled for presentation at the International Solid State Circuits Conference. "Less than four years ago, we embarked on an ambitious collaborative effort with Sony Group and Toshiba to create a highly-integrated microprocessor designed to overcome imminent transistor scaling, power and performance limitations in conventional technologies," said Dr. John E. Kelly III, senior vice president, IBM. "Today, we're revealing just a sampling of what we believe makes the innovative Cell processor a premiere open platform for next-generation computing and entertainment products." "Massive and rich content, like multi-channel HD broadcasting programs as well as mega-pixel digital still/movie images captured by high-resolution CCD/CMOS imagers, require huge amount of media processing in real-time. In the future, all forms of digital content will be converged and fused onto the broadband network, and will start to explode," said Ken Kutaragi, executive deputy president and COO, Sony Corporation, and president and Group CEO, Sony Computer Entertainment Inc. "To access and/or browse sea of content freely in real-time, more sophisticated GUI within the 3D world will become the 'key' in the future. Current PC architecture is nearing its limits, in both processing power and bus bandwidth, for handling such rich applications." "The progressive breakdown of barriers between personal computers and digital consumer electronics requires dramatic enhancements in the capabilities and performance of consumer electronics. The Cell processor meets these requirements with a multi-processor architecture/design and a structure able to support high-level media processing. Development of this unsurpassed, high-performance processor is well under way, carried forward by dedicated teamwork and state-of-the-art expertise from Toshiba, Sony Group and IBM," said Mr. Masashi Muromachi, Corporate Vice President of Toshiba Corporation and President & CEO of Toshiba's Semiconductor Company. "Today's announcement shows the substantial progress that has been made in this joint program. Cell will substantially enhance the performance of broadband-empowered consumer applications, raise the user-friendliness of services realized through these applications, and facilitate the use of information-rich media and comm

Re:Cell in TV ?

[bhima]

IBM is planning to market many different types of Cell CPU's for handys, TV's, Workstations, and and Super Computers.

Re:And this article tells us WHAT?!!

[TheRaven64]

The article on El Reg has a bit more information content. The chip is POWER-based, and supports multiple cores, each of which can run a separate OS. This is the first POWER chip to be produced in volume (I'm not counting workstation / server chips as volume). This, combined with the PowerPC-based XBox2 may mean that the unit cost of POWER/PowerPC chips drops enough to make beige-box POWER/PowerPC systems cheap enough to be a viable alternative to x86.

Small problem

[Craig Ringer]

There's only one small problem with your contention - Linux /does/ scale out to at least 512 processors - hardly 'a few' - and is heading up to multiple thousands with SGI's current work.

Of course, one could argue that the Linux folks have more than one insane genius among them...

Re:Why does a game console need such a monster CPU

[tuffy]

After playing games on a PC with a mouse, playing on a console is frustrating, not fun. And, I can not imagine playing a game that puts console based players up against PC players (unless you handicap the PC's hardware). The console players would be out of luck.

Playing first-person shooters on a console is frustrating, since the control is designed for a keyboard and mouse. Similarly, playing a typical console 3D platformer like Wind Waker on a PC will be equally frustrating since the control is designed for console pads.

It's the nature of the beast. And until someone sorts out the control differences, PC vs. console play will always fall short for one of the two.

Re:What language do they use?

[sleepingsquirrel]

Linux Insider is running a couple of editorials speculating about running Linux on the 'Cell' processor. The bold prediction? 'the Linux developer community will, virtually en masse, abandon the x86 in favor of the new machine.'

Believe the truth

[Ideaphile]

Sony originally promised the PS2 could render 75M simple polygons per second, but also said the geometry engine's limit was 36M polygons per second. This figure is accurate, but like all such numbers in the graphics industry, it is achievable only in a single-function demo app. Such figures are useful only for comparing the raw performance of different designs.

Sony never claimed the PS2 could support HDTV resolution. The company was very clear about the limited frame-buffer memory on the Graphics Synthesizer chip.

Sony did, in fact, make a multiprocessor PS2-based workstation, the GSCube, which combined 16 complete PS2-compatible subsystems. The "Graphics Synthesizer I-32" chip used in this system had a 32MB HDTV-sized frame buffer, leading me to speculate at the time (August 2000) that Sony would soon introduce an HD-capable PS2. They could have, but they never did. I can't get 'em all right, I guess. They demonstrated to me (personally) this system rendering scenes from the Final Fantasy movie in real time, so that wasn't hype either.

Here are the Microprocessor Report articles I wrote at the time (subscribers only):

http://www.mdronline.com/mpr/h/19990419/130501.htm l

http://www.mdronline.com/mpr/h/2000/0821/143402.ht ml

. png

Re:What language do they use?

[HidingMyName]

First a fwe points both regarding this post and the grand parent.

1. Multicell processors don't require processes/threads to cooperate/communicate via shared memory. What they do is permit a cheap version of SMP by packaging multiple processor cores in a single chip. Given the complexities of dynamic scheduling (I think an exponential number of gates may be required per stalled instruction that is tolerated without stalling the instruction stream). They can allow these cores to stall and cores that are not stalled can use the pins on the chip to signal the rest of the system. Much like on an SMP machine, not all jobs have to be parallelized, if you keep enough jobs running they can be farmed out over the processor/cores and the system will put them to use. Thus, I don't think multi-core/cell approaches will need more sophisticated languages/compiler/code generation techniques that are employed in VLIW, Simultaneous Multithreading/Hyperthreading environments.
2. The cores will contend for access to the system bus, since the pinout will not scale with the number of cores used. This could be tricky and form a bottleneck. The run time O/S scheduler may need to be smartened up to handle this (or maybe I was wrong above and some smart compile time techniques can help, but I'm not sure how that would work).
3. I don't think X86 is dead yest (at least not for the 64 bit variants). They fix some of the serious security issues (e.g. separate execute and write permissions bits for pages), and have a combination of a large amount of market share (hence big economy of scale) and a large amount of working deployed software. People aren't likely to walk away from that (yet). This may mean that early approaches could require X86 emulation to be adopted, which probabably means that speedups by a factor of 10 or more are going to be needed if software emulation approaches are used. It will be tough to get that.

Re:And this article tells us WHAT?!!

[TheRaven64]

POWER is not the same as PowerPC. While the PowerPC 970 inherits a lot from the POWER4+ architecture, they are different chips with different (although overlapping) instruction sets. POWER chips traditionally belong in high end workstations, servers and (I think) mainframes. PowerPCs are used in embedded devices (e.g. a large number of mobile 'phones) and a few computers (e.g. Macs, PegasOS sytstems, BriQs and the GameCube).

Until recently the POWER instruction set was a superset of the PowerPC instruction set, allowing PowerPC code to be run unmodified on POWER systems (assuming ABI compatibility). Motorola's 74xx series and IBM's 970 series added AltiVec/VMS/Velocity Engine, which may or may not appear in future POWER chips.