Slashdot Mirror
Cell Workstations in 2005
yerdaddie writes "The cell processor will be introduced in graphics workstations before release in the Playstation 3, according to press releases by IBM and Sony. As previously discussed, IBM will be releasing more details in February 2005. However, apparently prototype workstations have already been "powered-on" and will be available in 2005. Since Windows on PPC was scrapped back in 1997, this leads to speculation that perhaps Linux, AIX, or BSD will be the operating system for cell workstations."
I may be wrong, but to me this sounds like hyper threading with a new name. Can anybody enlighten me?
Electrons are free; it is moving them that becomes expensive.
Actually it sound more like parallel processing to me, where many CPUs are connected together to form one larger CPU. Perhaps you can remover CPUs from the network while active?? Or maybe it is just easier to expand. Their page seems to be full of hype (in my opinion), but no description of concrete benefits from this technology. Also why is this in the games section ... seem more like hardware to me.
Philosophy.
This article provides some background.
Well, knowing IBM and Sony there is a great change that it will run linux.
At the moment it seems that linux is the choice for development on the PS2 and I think it will be with the PS3.
I wonder what the average speed of the processors would be? And if they'd include HyperThreading?
It has been stated before that the PlayStation 3 is expected to be capable of distributed processing due to the capabilities of the Cell architecture. Whether or not that will indeed be the case remains to be seen, it is certainly a lofty goal for the current market penetration (not to mention speeds) of broadband in the home. Does Sony expect these PS3s to cooperate with their Cell-based television sets?
.sig: Open Source, Open Mind
For all practical purposes, the PowerPC has been relegated to a Mac-only solution while high performance NT users have turned to Digital's Alpha....
This move puts Apple Computer in another awkward position: the company had been planning on using Windows NT in its Web servers.
And my favorite actual fact is that microsoft is going back to Power PC with the new Xbox . But Im sorry that Alpha has been erased from the map.
"We all know Linux is great...it does infinite loops in 5 seconds." -- Linus
The graphic
Nothing to see here
The article says that each chip is running its own kernel. That seems like a lot of wasted energy to me. I agree that it could give a serious boost to performance. However what about the memory requirements (RAM specifically)? It sees to me that each micro-kernel is going to need some RAM of its own, and to get the promised performance you would need many of these micro-kernels. This technology may end up more limited by memory requirements than the speed of the chips.
Philosophy.
running a atandardised obtainable kernel(windows, linux, ect), with hardware level access after deencryption (if it's even encrypteD) and an exploit
.....
Am I the only one here thinking "bad fucking idea" or what? And lets not even mention the latency for distributed supercomputing applications. Everyone is now on wireless, unsecured, and sending signals all over the place. Hell, I should support that; free internet with the touch of a button after hijacking someone's toaster. w00t.
Candy-Coated Knowledge
No. You are mistaken. Xbox will have a PowerPC derivative. Not a POWER derivative. Also, I should note that Cell, although part is derived from the POWER4, is not really anything like the POWER4 architecture.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
The most interesting part, however, is that MS may be putting up .NET as the development environment for the X-Box 2. It makes sense that MS would try to leverage their gaming platform to lure developers onto the .NET platform and commit their engines to that API.
On another note, could Linux and Mono play much of a role in this if the Cell does indeed provide a Linux environment for development? If Sony is able to provide a less expensive development environment, development costs may ultimately go down and the consumer would benefit.
This could be either by the increase of choice since the bar of entry would be lowered for smaller software houses, or by cost if the games are indeed cheaper as a result; Existing engines and software could be ported or would be compatible, or due to the the ease of coding on a familiar platform.
.sig: Open Source, Open Mind
From one of TFAs: The Cell workstation is designed to deliver tremendous computational power, helping digital entertainment content creators generate higher quality content with richer and more dynamic scenes, much faster than current development systems.
This points at more than just game consoles. This looks like Sony is looking ahead to a future in which they can dispense with actors entirely and rely on realistic computer generated characters. Should be a good bit of money to be saved if you don't have to pay an actor millions to star in your film. Could be other applications too: Animated news announcers with features finely tuned to inspire trust in the viewer, human-like avatars in intelligent appliances, human-like answering machines and customer service line responders, etc.
So, how far are we from the footage ala William Gibson's Pattern Recognition and the "live" entertainment ala Neal Stephenson's Diamond Age?
Bureaucracy loves company.
You'd boot into something like Grub and choose your processor. That way you could run a UltraSPARC workstation, MIPS, Itanium, or something as small as a PIC. It'd be great for cross-platform development especially for embedded users.
I'm sure processor hobbyists would spring up to fill every niche of emulator. Probably be a great proving ground for design theory.
Considering the low heat output you could have a dual/quad-processor box.
Maybe someone would figure out how to run multiple translators at the same time so you could run x86 and PPC and 68K at damn-near native speeds
To me that'd be the ultimate workstation.
Your CPU is not doing anything else, at least do something.
The more of these you have in your house, the faster the game/app you're playing/using will run as it will automatically use spare capacity on the other machines networked together in your house... I for one am most certainly looking forward to getting my hands dirty coding for these beauties... Bring on the Cell Processing Overlords... I'm ready.
Donald 'Duck' Dunn: We had a band powerful enough to turn goat piss into gasoline.
In all fairness, we really don't know from what processors the CPU in Xbox2 nor Cell will be derived from, but the most reliable information we have says that the Xbox2-CPU will we quite similar to PowerPC 970, but with three cores. As such it is indeed a PowerPC processor but it is also POWER4 derived. Cell on the other hand is stated to have a 64 bit Power core, and that's quite different from saying that it's POWER4 derived. IBM uses the term "Power" for both PowerPC- and POWER-processors, so it very well could be, and probably is, PowerPC-based and not POWER-based.
The core in Cell is probably an highly evolved PowerPC 440 based core since that is a quite proven, capable, lean and have a very modular design. I think it would be unwise to build Cell around a massively complex design like POWER4. It would suffer immensely from compelxity, power consumption and its monolithic design.
- Henrik
- when the Shadows descend -
How is Windows on PPC even relevant?
IBM aren't big on windows anyway, or has the endless IBM/Linux advertising passed the world by?
I think we can be pretty confident the OS *won't* be Windows.
I'm still wondering about the real-time uses of this, i.e. PS3. Latency becomes a huge issue when you're trying to render a frame every 16ms.
ISSCC show information that has leaked show that Cell includes a POWER core of some sort. It seems to be the manager of all the stream processors. Beyond3d.com has a good forum thread on this. You are right in saying that POWER can be PowerPC but it usually doesn't. As far as XBOX2 goes, the PowerPC 970 or later derivatives would be the prime candidate as the developer kits are basic PowerMac systems running Microsoft software (ironic). I don't think Cell will be ready for mass production by the time XBOX2 will be realesed so even if MS wanted to pay good $$ for it, it probably wouldn't happen.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
While PPC support was dropped, if I recall correctly back in the Win NT 4.0 days, NT was amazing because it was designed from the ground up because it could basically be compiled for any endian chip/any aritecture.
Since it is the core of the current and future lines of windows, the windows base should be portable to a cell based system, basically it requires some new drivers and probably tweaking of the HAL abit. The problem is that all the applications (that we all consider part of the windows os but are really just applications running on top) would need to be redone.
Microsoft would have one of these machines in house by now for they're windows teams to work on supporting. That I have no doubt, what I do doubt if microsoft will consider this important/the future and if they'll support it during the inital release (w/ longhorn maybe?) or if they'll come late and lose a large section of the market as we all jump and have to use a *nix as the desktop.
If this whole cell thing is more then hype, and is the wave of the future, Microsoft will support it.
The same processor powering cell phone, PDA, gaming handhel device, gaming console and general porpose workstation can be a way out of porting-emulators hell which is handheld development is for now. However there will be different OS for handhelds still probably - for examle Nokia unlikely drop Symbian in favor of Linux...
Yes they are significant. Never underestimate the power of money. I probably won't buy an XBOX2 because of XBOX. I will buy one because of good titles. Disclaimer: I own all three major console right now and I play XBOX the least. I play it when I can get a better game or an exclusive game (Halo1/2) but I still play my inferior PS2 and GCN because I like the games (Katamari Damacy, Paper Mario, ICO, Ratchet & Clank, MGS3, F-Zero GX, etc...) To be fair, XBOX and MS aren't all that bad of a console. They are not better but not worse either. It is also a personal thing. There are many people who would probably find XBOX has the better games for their taste.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
Yeah I can see it now. New hot technology platform released with lightning fast speeds, and broad industry appeal. Yep, thats somthing Microsoft won't touch with a 6 foot long stick. Please... if Cell proves to be a useful and profitable platform, Microsoft will move in.
What'd I'd like to know is what IBM's solution to the software problem is. Software has always been the achilles heel of multiprocessor systems. Most existing programs and even most existing programmers can't use the resources efficiently. That's why we have gargantuan superscaler, out of order processors. Expensive in terms of hardware but it suits the software better.
So, why is Cell going to be easy to program, when other parallel systems aren't? The bits of that i've seen about the architecure suggests that programming might be an absolute bear.
Well, you are wrong - PowerPC is based on POWER architecture and any software runs on PowerPC will run on POWER4/POWER5 chips.
I am not talking about intruction sets. I am talking about architecture. The POWER4 is a different line than the PowerPC. What you say is like compairing the Pentium M and the Pentium 4. The P4 can run PM programs easily but they aren't nearly the same.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
Uh... We are talking Cell not XBOX2 SDKs ok? If otherwise please post your sources. I would be interested in proof as I have not seen ANY reason to think Sony shipped PowerMac G5 workstations to its developers.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
the Macintosh OS might be an easy port to to PPC...
All POWER-processors have been fully compliant (32 and 64 bit) PowerPCs since POWER3, and before that the RS64-procesosrs were too fully PowerPC compliant. So.. you are wrong in saying that most POWER-processors isn't PowerPCs since they have been since 1999, and they have been even more PowerPCs than "clean" PowerPCs since they until the 970 didn't have the full 64 bit ISA.
:)
The ISSCC papers state that Cell is Power based, not POWER based. There's a significant difference here since IBM in its marketing use the "Power" moniker to encompass both PowerPC and POWER processors. If you have seen different papers than I have, please provide me with an URL of PDF that proves me wrong. This is important stuff
- Henrik
- when the Shadows descend -
You are really slow. Really. Read the parent. Then read my reply. The "latest" powermacs are G5 based. Apple has not even hinted at adopting such an architecture. You are the slow one. Come on. Read the article for once. Learn what Cell is then come back and post please. One last time: Read the whole thread.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
I think IBM, Toshiba and Sony eventually will license and sell Cell technology to those who are interessted. One of the core ideas is that they want to spead this technology as far at they can since every Cell based machine tap on the computational power from all other Cell based appliances in its vicinity. The more the merrier!
Cell isn't one processor, it's a class of processors. The one that will go into the workstation is more powerful than one that will fit into a PDA, or a HDTV. I think that IBM will make one workstation, and Sony will make another. They will use different boxes and logos but they probably will use a common "Cell based" lable yet unseen, just like "Intel Inside".
- Henrik
- when the Shadows descend -
Although I can't find a link now, this reminds me of a concept I read IBM touting a couple of years ago.
Aimed more at the "enterprise space", this article talked about cube shaped modules or nodes, with an I/O interface on each of it's six sides.
These nodes would each snap together to form a larger array, like Lego bricks. Each node would have an specific purpose, such as processor node, or data storage node, or whatever else may be required. Every node could communicate with every other node.
Managing a data centre would become like growing a biological organism as the system expanded. You would add new nodes to the "front" so the system would move around the computer room floor. Old dead or redundant nodes would emerge from the back.
I don't know how they planned to deal with latency across the nodes as the system grows, but still a fascinating concept.
Serving Suggestion: Defrost
I have this info scattered through my posts. What I am trying to say is that an ISA does NOT make a processor the "same". You seem to be intelligent. I am sure you can figure the rest of my points out with this in mind. Many good URLs were posted to the Beyond3d.com forums.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
The development kit for Xbox 2 is Windows NT4 for PPC with Xbox 2 extras.
In essence Cell is just that, but it doesn't stay there. Cell technology can distribute it's load to other Cell processosrs nearby. It's built from the ground up to use grid technology transparently. Quite revolutionairy.
- Henrik
- when the Shadows descend -
No. Cell processors are actually cheaper to produce than x86 based processors (less transistors and logic wasted on bandaid solutions like caches and deep pipelines). One of their strong points.
Hmm. I like my XB much more than my PS2. Yeah, there are more games for the PS2 -- but most of the great titles come out for all platforms and the XB ones usually take advantage of the better graphics and HDTV capability of the XB.
The XB is also much more useful once modded than the PS2 could dream of being. You basically end up with a cheap and functional HD media player. Add an extra HDD, use as an mp3 player, install Linux, watch movies, play around with homebrew software, install the mame archive, use XBC to play online with friends without using M$'s service, do whatever!
I find it both sad and hilarious people will make themselves hate a great device just because they are unable to think for themselves and must jump on every popular bandwagon that comes along.
I'm sure some compulsive anti-microsoft flame artist will reply and tell me how evil they are and name every offense they have commited. Guess what? I don't care! Did everyone forget Sony is a huge player in the music industry? You know, that bunch of freaks who have sued the pants off of thousands upon thousands of our fellow civilians?
I'll leave you guys to argue amongst yourselves as to how half the GPU, CPU, and RAM with analog video output makes for a superior product at the same price as the alternative.
Well, they had to take their idea somewhere when their glans desktop never made it off the drawing board.
It would be cool if it didn't suck.
Actually, CELL is based around the 970. Expect about 80-90% performance compared to an equivalently clocked 970. Where it goes nuts is that there's a number of vector units attached that are basically "VMX on steroids" to quote one of the main guys at IBM behind this. The vector units (or Data Plane Processors as they're calling them) can also communicate between each other as well as with the central core. The workstations are actually headless server blades, each of which will have 2 CELL's on them and they'll be running Linux.
This stuff isn't bullshit, it was all disclosed Thursday at the Australian Game Developers Conference. I didn't sign a NDA so it's all good. I also fondled a PSP =]
Computer components that talk by wireless..
;-)
That must be the wet dreams of NSA employees
I'm still trying to figure out what people mean by 'social skills' here.
Cell workstations will be 8-way tipically, which many programs (like GCC) is able to use. If claims of Power5/Cell performance are true, it means that it will compile linux kernel under 5 sec. (8-way). All system, including KDE/GNOME and standard set of apps will take less then hour. Sounds too cool to be true.
839*929
You know.. If we agrue this much about it: It doesn't matter. It can be seen either way so lets just drop it and let the readers decide.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
There are two problems with this post. The first being that is anonymous, which shows cowardice. Perhaps this is because a vital piece of information is incorrect? Microsoft had no real trouble putting Windows in AMD64, but getting all the driver venders to support it is taking a long time. For Linux where all the included drivers likely have source or are set to run in 32bit mode this is fine. However, would you want your drivers to drop back to 32bit? Probably not... Which is a likely reason why MS has not let Windows XP 64bit Edition loose.
-]Phreak Out[-
XBOX2 will be based on parallelized PowerPC G5 processors. It is not Cell tech. Game dev studios are using PowerPC G5 macs to develop next gen titles for the XBOX.
There are rarely new ideas in computer architecture. This may be an update of a transputer-like thing with better communications and core cpu. The real questions that need answering are about the memory model (distributed?) and comm model (warmed over message-passing?).
Chances are it contains nothing new or radical
and is likely just a cluster component on a chip.
Prepare to be disappointed.
I'm still a bit worried that I've not heard much about the seemingly built-in DRM management of this new platform (that seem to be able to spread to all facets of technology, including toasters). According to a clause in the pressrelease by IBM and Sony from Nov. 29, the Cell processor will have:
- On-chip hardware in support of security system for intellectual property protection.
Is this the end of tampering-capable hardware (e.g. machines where you can modify the kernel, bypass DRM-systems etc) that some people have long foreseen? Anyone more in-the-meat of the technical details care to elaborate on this?
avocade.com
In a free and open internet, who needs Windows
Nothing's official just yet, but this is WAY more interesting than studying for finals, so here we go:
Processor instructions are broken into an 'apulet', which contains data as well as code to perform an operation. This is probably why its claimed that if more processing power is needed, then its a simple task to add a new workstation and the work can be offloaded.
A cursory read suggests that its like creating a cluster of highly efficient yet simple nodes.
Corrections are welcome.
Reference: EETimes
Windows development for PPC is most likely well on its way, as the next XBOX is PPC powered as wel. The development kit for the XBOX 2 games runs only on Apple G5 systems, and rumors state the next XBOX will be availeble in three versions, standart/with HDD etc/full PC, so it looks like Microsoft is supporting the PPC as a workstation processor as well. I'm quite exited to see what the next two years will bring
Interesting thing is, there will be several variations of Cell processor in its lifetime - PU can be any Power ISA processors, or something different, after all. A high-end Cell may use Power6 while a low-end Cell may use PPC440. That depends on usage and required spec.
Interesting stuff. Any more details you recall from the conference? In particular, how many vector units per 970?
++ Say to Elrond "Hello.".
Elrond says "No.". Elrond gives you some lunch.
Cheers for the info shplorb! Any more details? Did they disclose how many "Data Plane Processors" were attached to each 970? Any mention of flops performance? Clockspeed? *Anything* else? :)
Many thanks..
FPGAs can do what you're suggesting. Check out opencores for some examples of processor implementations already written. Assuming a large enough FPGA, you could implement any processor out there (and throw an Ethernet controller, VGA controller, etc onto the same chip).
That said, there's a huge performance difference between a real processor and a "soft" processor. The soft processors can't be clocked anywhere near as fast as an Intel/AMD chip. I think Xilinx is boasting a 200MHz Microblaze on their newest FPGAs.
It's a lot more advanced than that, the cores don't need to be on the same board, they can be seperated by a network connection.
It's system on a chip architecture and it's a lot more elegant than anything Intel or AMD will come up with, simply because it is free of x86 compatibility.
I actually am against mods in general. Too many people pirate (they claim they don't -- but you know they do). I am for homebrew (which I do some myself). Mods on XBOXs give very little feature wise that is IMHO clearly on the legal side. (Don't flame on this please... I just don't think it's for any good unless you are a homebrew hacker or a curious fellow -- I know most of the world doesn't fit in that catagory yet there are still so many mods.) I also son't think graphics are the only reason to get the XBOX version. Many people like XBOX Live (I don't want to pay ;) ). I hate the XBOX controllers (both large and small)... In fact I like the game cube's best for many games while d-pad centric games are great on the DS2. But some people like the XBOX controllers for their hold style.
People are too caught up in very little difference in video quality. Its a GAME. play it. It's not hollywood here. Gameplay has always come first for me. Don't get me wrong: graphics do matter, up to the point where you can express what you need to.. . anything beyond that is just eye candy. Eye candy is nice but I will forget about it after I start thinking, "DAAANG! This is FUN."
As for modded functionality: I can do it on my PS2 (or my XBOX if I modded) with the independence exploit OR the Linux kit. I use both depending on what I am going to do.
Things I do like about the XBOX: Four controller ports and games actually use the HDD. I am not really the kind of guy who even has enough money to own an HDTV, so graphically, games are close enough or good enough for me.
Things I like about the PS2: Fun (and hard) to program (it's a challenge), some GREAT games (no I do NOT like GTA :) ), good controller design, and plays my PS1 games.
Things I like about the GCN: Fun games with some good exclusives (yes they exist), four ports, cheap, easy to run imports (freeloader).
There are some things I like about all of them in general: All are capable of quality graphics (Progrssive, good 3d, etc...), games (again :) ), and good multiplayer (the human factor makes things interesting).
There are some things I don't like about all of them but I will not start a flame war over this off topic post anyway :)
Point: It doesn't matter. None of it does. blah > blah > blah is just junior high kid talk. Get what you want and don't let anyone tell you that you won't have fun because: You can have fun on all of them. One could liken this to how people follow sports:
1) X team sucks!
2) No they rock!
1) No Y team could beat them any day!
2) Y team has no clue what they are doing!
Narrator/me) Neither team "sucks". They are both professionals. one may win a game but that doesn't take away the fact that the other still knows how to play well. It may not be as well as the other team but it's still darn good.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
Good point. I would mod you up if I could.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
Interesting. Thanks for the post :)
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
If I recall correctly, Sony Playstation 2 workstation (the one with the emotion engine) was over 15,000 USD. That puts it well beyond the "that would be interesting" price range and most likely beyond the aspiring game producer just out of college types.
Where, I would hope, a PCI-X based card could probably be priced much lower.
Now that I've said all of that, the old workstation would make an interesting addition to my collection...
Nothing in the world is more dangerous than sincere ignorance and conscientious stupidity.
Making a distinction nowadays between Power and PPC is like saying "Ahah your OS may run on Pentium 4 CPU's, but will it run on Pentium 3 cpus?"
Well , that really depends on whether or not the OS in question utilises processor extensions and optimisations only found on the P4.Consider the difference Altivec makes to PPC's for example. While the os may run on both architectures there will be noticable speed differences and improvements even though both chips have a similar heritage. You shouldnt assume - particularly in the case of game consoles that these chips will be running a generic OS/Kernel that will run on multiple different incarnations of a specific processor family. It is far, far more likely that it will be tailored specifically to take advantage of the wonders that the Cell processor used in PS3 has to offer, and that is when the differences between a Power and Cell really start to matter.
Electronic Music Made Using Linux http://soundcloud.com/polyp
The POWER train seems to be in full motion. No more wondering why IBM is canning its x86 desktop crap.
I infer this means a full shift into Power based architecture from IBM, they will only retain x86 server products because customers may want them, but they will not play a large role in their roadmap.
And that could be a Very Good Thing. The Power architecture is superior to all x86 implementations, including AMD64, in every way. The sooner we can break out into full uncrippled 64 bit computing the better.
From http://en.wikipedia.org/wiki/Transputer
This seems like an excellent opportunity for Apple to license Mac OS X.
I'm assuming the intruction set for the cell processor is a superset of the existing PowerPC processors, or that the missing instructions could easily be emulated. If so that would make this is a graphics workstation that could run Photoshop, Final Cut Pro, Shake, and other top notch professional software immediately. The existing user base wouldn't have to buy new versions -- their old versions would run.
As discussed many times on slashdot and elsewhere, Apple won't license their OS unless they believe they can do it without cannibalizing their existing user base. Doubtless there would be some cannibalization of the high end, but if it makes OS X the clear platform for high-end graphics workstations it could still be an overall boost to Apple. I don't really know how the current high-end graphics market sees OS X. My impression is that a surprising amount of it is on Windows, and that Apple is just holding on to its market share in this area.
Anyone with more current knowledge of the high-end graphics market care to comment?
As a game developer working on a Windows product that needs to be portable to XBOX, I can assure you that XBOX != Windows.
First, the XBOX supports only supports APIs, such as DirectX, widely used in games. It doesn't even come close to supporting a majority of Windows API calls. And it doesn't support DirectX quite like Windows does. It suppots a superset of DirectX 8, but not everything in DirectX 9.
Graphics apps are the most likely of all non-game apps to use DirectX, but they are likely to use many of the other API calls as well.
In addition, even if a particular app you want to run is fully supported by XBOX (in terms of API calls), the hardware and those calls are aimed at and optimized for games (ie: expect game controllers, not mouse and keyboard).
So I think Microsoft would have a fair amount of work ahead of them if they were to try to bring the full blown Windows OS to a new platform with a cell CPU. How hard depends on how much they've maintained the CPU portability they had with Window NT.
I was there also. They did not disclose the number of DPPs on each unit.
there's a number of vector units attached that are basically "VMX on steroids"
Did they offer any details on how folks are going to program these beasts? Will you have to write assembler to take advantage of the vector units?
Also, will they release their own compiler, or port gcc? There's been a fair amount of traffic on auto-vectorization on the gcc list over the last year or so, but while I've seen a few Apple people there, I don't recall seeing anyone from IBM. (Or Sony, for that matter.)
Wife: Honey, can you turn down the TV volume, you're stealing too much processing from the microwave and my chicken wont bake nicely.
Husband (sniggers): Yah, as if it'll make it taste better
Carpe Diem: Seize The Day!
It is being used to program games for the XBOX next, via dual G5 machines from Apple. I doubt the games are programmed on MacosX,plus, more than one source seems to agree...
but most of the great titles come out for all platforms and the XB ones usually take advantage of the better graphics and HDTV capability of the XB.
eh, doesn't the PS2 have HDTV capability as well?
-mkb
Too bad 3M didn't get involved.
Then it would have been the STIM Cell processor.
isn't the ps3 gonna be back compatible with the ps2 and ps1. if so the cell would run on a mips r3000
instruction set and not a ppc instruction set.
any one knowing anything about this????.....
Time to move to Canada.
Take off every sig. For great justice.
Technological Features for "first-generation" Cell chips:
4.6Ghz Clock Speed
1.3V operation
85 degree C operation with heatsink
6.4Gb/s off chip communication
from the article:
eight cores on a single chip
90nm SOI process
Link to Powerpoint
Link to Original Article in Japanese
By chosing from 2 threads instead of one it has greater chances of finding an instruction that can be computed by an idle (at that time) unit
If the computer has, say, 100 K or 1 M cells running at low power levels, it can still compute a difficult sequential problem with a speculative approach much like the hyperthreading scaled up to the stratosphere. Variables that take a long time to compute can still be assigned speculative values and used to compute further. Then when the values of the variables are definitively obtained, the results based on incorrect speculative assignment are discarded while the results based on correct assignment are retained.
Analogies to this approach occur readily in theorem proving. Fermat's last theorem was shown to be equivalent to a number of conjectures. A fully sequential approach would have to prove the conjecture before proving the equivalence while a speculative approach allows the proof of the equivalence before the proof of the conjecture.
The analogy breaks down in that the conjectures had to be equivalent to Fermat's last theorem rather than a larger set of assertions with unknown equivalence. Another analogy is one might have used Fermat's last theorem speculatively before it was proved, to derive results that would be accepted only after Fermat's last theorem was verified.
Willy nilly speculation would waste a lot of processing time (but so many computers are devoted to nothing but speculation anyway). There are two coordinates - time and space. The speed of light constrains time, but space might be used to obtain results faster than pure sequential computing.
Speculation on major variables that have only a few possible values is practical - a lot better than speculating on a large set of variables that imply an exponential number of speculations.
Know your pads. One time pad: good for cryptography. Two timing pad: where to take your mistress.
As an example, the PowerPC-AS series doesn't support 32 bit PowerPC instructions. You will not get an out-of-the-box runs-on-your-G3 Debian installation to run on it.
The Power family is a family because they're all related, but just like real families, there are a lot of siblings in it who are far from compatable, have wildly different interests and strengths, and who will not deal with the same friends.
You are not alone. This is not normal. None of this is normal.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
According to
m l
http://www.linuxinsider.com/story/34994.html
longhorn is not going to run on the cell chip:
"Microsoft's software can't take x86 beyond some minor hyperthreading on two cores without major reworking -- and Itanium simply doesn't cut it. The Wintel oligopoly could spring a surprise -- a multicore CPU made up from the Risc-like core at Xeon's heart, along with a completely rewritten Longhorn kernel to use it. But no one has reported them stuffing this rabbit into their hat. So, for now at least, they seem pretty much dead ended."
This article predicts that the cell chip will replace x86 as the main platform for Linux.
http://www.linuxinsider.com/perl/story/34707.ht
Each processor runs it's own kernel. I think you better think of a real micro kernel. And the features called where the instuctions are close to the data sounds more like some kind of message passing kernel like the mach-kernel.
It would not be a problem to run a GNU/Linux kernel on top of that.
But forget the 1 memory - x processors pardigma. That is not needed for a system like this and would only creat a bottleneck.
Not that that answers the question, but AltiVec is not a prequesisite to run any existing version of OS X.
You are not alone. This is not normal. None of this is normal.
I want a "cell workstation" that's a mobile phone with UXGA display (HMD? Hologram? Neural interface?), voice recognition and 1.5Mbps Net connection. Why must IBM bait and switch me with this deskbound processor terminology? At least they could roll out a WiFi microfluidics DNA nP.
--
make install -not war
You make some very interesting points. I would also like to add that the XBox 2 development kit is a PowerMac G5 (PPC970fx) with Windows (NT?) running on it.
With Microsoft choosing PPC processors for it's XBox 2, and Sony choosing PPC for a new class of workstation, things don't seem quite as rosey for the x86 architecture as they used to.
"Politicians find new names for institutions which under old names have become odious to the people."
An XB is $149 and comes with two games and a controller. The $149 will cover the cost of a case and powersupply. Now what?
Not really. You can purchase a component kit, but it will only work on the SCPH-50001 series PS2 and newer. There are like 5-6 games that support it, most are not even in 16:9, just progressive scan.
Well said. I think a lot of people watch ET and hear about the huge salaries and think, damn, all those actors are overpaid! What they don't realize is that about 1% of all actors make enough from acting to live on.
More to the point, it's not as if acting is the biggest expense on a movie. Most movies, the film stock alone costs more than most of the actors. When a film does have a huge actor salary, it's for a reason. The producers sign Julia Roberts for $20million because they know that her name alone will make them more than that at the box office. So bringing in a whole CGI team to replace the actors doesn't exactly sound like a cost-effective measure to me, since you'd need a few people plus a lot of equipment to do the work of one actor. I'm not even going to get into what acting actually involves and how you can't just program it, because if I have to argue that point with anyone, it's a waste of time.
I'm not a big windows fan. Nevertheless, what I do know about the architecture of Windows NT (yes, that includes 2k and XP - they're just newer versions of NT for all the marketing hype) suggests that porting it to a new architecture should not be that different from porting, say, Darwin.
0. We presume there's already a toolchain available to generate binaries from C.
1. Write a new HAL. The HAL is the Hardware Abstraction Layer at the bottom of the NT kernel. It is the interface to a bunch of very low level services, like memory management and basic I/O. Probably more things I can't think of right now. It's important, but relatively straightforward.
2. If this new machine doesn't, more or less, have a PCI bus for most of the major peripherals, then there will be a bunch of driver work at this point.
3. I wave my hands here, as I do not know all of the details of booting NT. You need to load the HAL and the kernel. I am not sure when control is handed to the kernel to load the rest of the drivers from the boot media. But the boot procedure will need to at least load the HAL, the kernel and at least a few drivers. If the new machine uses OpenFirmware, then this will likely involve a primitive NTFS filesystem reader in FORTH. Such a thing may, in fact, already exist.
4. Compile the source. Presuming the machine is either LP64 or ILP32, there ought not to be any real difficulties, depending on how together Microsoft has their shit. They've been down those two roads already. I am less sure about endian issues.
5. ??? (usually involves violations of anti-trust laws)
6. Profit!
Extending it to 64-bits may be against the licensing agreement that they have with ARM. Or if it wasn't, they may have to turn that IP over to ARM at a later date. I don't think this will ever happen.
2. Alpha reborn. Less likely but it is know to scale well to high speed and smp. Could the Alpha scale down to PDA level?
I love the Alpha architecture. It is a very clean RISC architecture. However, instruction density is a little low.
3. Forget RISC. Maybe RISC is not the way to go after all. The difference between memory speed and cpu speed is increasing. I code density the key to higher speed in this day and age. Should we think about super CISC where each instruction does more?
Ding, ding ding! We have a winner. Both Alpha and ARM (XScale) are RISC architectures. Itanium, Alpha, PPC, and (to a much lesser extent) ARM, have instruction density problems (they require larger instruction caches to get the same effenciency as an x86 chip). This problem only gets worse when CPU core speed keeps increasing at a faster rate than memory speed.
Even x86-64 has its problems. For a long time the x86 arch was seen to be register starved. When AMD release the x86-64 arch, it doubled the number of integer and SSE registers. However, to make use of the extra registers, you needed to add a byte to every instruction. This, combined with doubled pointer sizes, meant that many 64-bit programs run slower than 32-bit versions (especially if the compiler couldn't use the extra registers effectively). This is less of a problem with the Opteron due to it's integrated memory controller.
Anyway, the main reason I replied was because you simultaneously advise going for Alpha and super CISC. These are completely opposite directions. The RISC / CISC wars are funny. CISC was used to keep programs small (no space to store them). RISC pulled ahead when hard drives and memory was cheaper. CISC is taking the lead now because HW is sufficiently fast to decode it, and memory bandwidth into the processor is severely constrained. I don't see the trend reversing again any time soon (especially with multi-core architectures sharing memory bandwidth).
Dan
Wow... you mean people actually play and enjoy Katamari Damacy? Personally I think Greg Dean said it best...
"I thought my PS2 took a hit of acid or something."
- Just so we are clear, Power4 is a PowerPC chip
Eh, wrong.- I'm not sure exactly what distinction you are making...
Distinction:POWER4 is a POWER chip, with the full POWER ISA, running on an IBM bus. As are POWER, POWER2, POWER3 and POWER5.
ALL PowerPCs (601, 603/e, 604/e, 750/g3, 7400/750FX/7450/g4, 970/g5) include only (diffrent) subsets/supersets of the POWER ISA. Also, they use a Motorola bus.
In conclusion, wrong.
See: http://en.wikipedia.org/wiki/IBM_POWER
There comes a time in every man's life when he must say, "No mother! I do not want any more Jell-O!"
That's Ice Cube.
Yea this was truly sort of stream of thought. ARM really does do well with instruction density espicaly when using the Thumb ISA in the embedded space. The question still becomes a heat/power to complexity issue. We are rapidly running up to heat limitations on the CPU. The more complex the ISA the more heat the chip seems to through off. I mean really when you are going to liquid cooling something has got to give. The PowerPC and the Power4 line really seem to be moving ahead of the X86 on many fronts. I have to wonder if IBM might be selling off it's PC division as a prelude to a shift to all AMD64 and PowerPC systems. Dump Microsoft and Intel in one move.
Arm seems to have a lot of things going for it. You are right about Intel not pushing it out of the embedded space since it is NIH.
Frankly I do think that SuperCisc is the way to go long term but XScale and Alpha where/are solutions that Intel could do quickly and frankly it really needs something the Itanium is looking like a new 432i or 860i.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
I don't know that "The Cell" is the answer, but this kind of distributed CPU load has been a long time coming. Sitting at an ad agency watching eight dual G5 PowerMacs sit idle while two more slug through some Photoshop work makes it really apparent that what IBM / Sony are trying to accomplish is much needed.
Kick it up to a higher level, a 5,000 user corporation, and you could replace million dollar server farms with Cell workstations.
And the home appliances market could be super cool - the more Cells you have, the faster your pr0n server can comb through your nightly batch downloads (:
No sig for you. YOU GET NO SIG!
Instead of programming for a bunch of specialized CPUs (each with their own strengths and weaknesses), all the processors are the same. Even better, depending on how the OS works, you might not have to think about separate processors at all! Best case, the programmer just sees a single pool of available processing power.
The big advantage for hardware developers is that instead of spending $$$ to make the GPU fast enough to handle all the polygons of a graphic intensive game, the audio processor fast enough to handle a sound intensive game, and the "Emotion Engine" fast enough to handle an AI intensive game you just add as many Cell CPUs as you can and let the developer figure out how they want to use that power.
Spell cheek you've failed me four the last thyme!
Ironically /l stripped them out... I'll put
's in next time
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
Look. It doesn't matter now. I was trying to say that I was not thinking they sent those types for the same reason you flamed me. "reading comprehension" please?
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
I am sure the devs were on drugs... no the less, it is still VERY fun.
ruby -le"32.times{|y|print' '*(31-y),(0..y).map{|x|~y&x>0?'
They are better known as 'cubicles'
Okay, first off, the bus isn't really germaine to a discussion of the architecture. Further, what bus do PPC's like the 440 series use? Hint, Motorola didn't design it. 486's and Pentiums are the same architecture but they use a different bus...
h tmlc /
s g/ 2a09c59094e16205
Second, while the wikipedia article is interesting, I'd like to refer you to
http://www-03.ibm.com/technology/power/about.
https://www-03.ibm.com/chips/products/powerp
I think they have some familiarity with the subject. The first URL lists a bunch of PowerPC documents under "Power Architecture resources"
From the second URL, the quote, "As members of the IBM Power family of products that power everything from handhelds to mainframes, enterprise class servers to video game consoles, IBM PowerPC products deliver high performance, power optimization, and excellent value."
So, IBM sure seems to think that PowerPC stuff is a member of the POWER family. You better let them know that they goofed up.
Now, just to make things perfectly clear:
http://groups-beta.google.com/group/comp.arch/m
(hopefully deep linking is working right on the new groups - that should be a post by Mr. McCalpin himself from Oct 19 2002)
To quote : "POWER4 is an implementation of the 64-bit PowerPC architecture.
PowerPC 970 is an implementation of the 64-bit PowerPC architecture."
He worked on the POWER4, so I'd trust him.
twit.
And it's rather amusing to think that Microsoft had NT ported to a 64bit processor a long time prior to the introduction of the Opteron.
They never did port MS-Windows to 64-bit alpha; it only ran in 32-bit mode. Compaq was involved in the 64-bit port, but announced in 1999 that it was foregoing 64-bit development in favor of IA64.
Dave Cutler *did* get some early versions of 64-bit Win2k to boot on an AlphaServer, but since Compaq lost interest in developing Win2k for the Alpha (both 32-bit and 64-bit versions), MS decided to pull the plug on Win2k for alpha entirely.
This was right at the time (late fall of 1999) that Intel sent out the first of the Itanium chips.
Anyway, MS never did finish development on a 64-bit version of MS-Windows on Alpha.
Microsoft is to software what Budweiser is to beer.
I think Sony probably feels vindicated by the success of the Playstation2 and its wacky architecture that, lets face it, is probably contributing to untold man-centuries of additional work than if the xbox or GameCube had won the console wars. Of course since everyone spent all this time learning how to do all the l33t PS2 tricks everybody spits on the Xbox as being for wimps...
While the idea of coding an entire project in little streaming modules that run on a zillion little processors is certainly academically interesting, if we won't be allowed to run our "main logic" code on some kind of CPU with more than 128K of addressable memory, that will be a huge barrier to entry, and could put Sony at a huge disadvantage compared to more conventional consoles. At least with the PS2 you could run most of your code on a fairly normal CPU and just focus on putting the graphics pipeline on the vector units, and maybe a few other things as needed, which is a lot of "bang for the buck", plus it lets you maintain a clean, multi-platform structure.
Well, here's hoping that at least the APUs will have such "frivolous" features such as integer multiply and hopefully a C compiler... also, if the guys who wrote the rather bloated IOP "operating system" code are involved with this project, that could be nasty... at least with the PS2 you could go straight to the metal and not have to care too much that nobody at SCEI seems to know how to write proper code...
Since Windows on PPC was scrapped back in 1997, this leads to speculation that perhaps Linux, AIX, or BSD will be the operating system for cell workstations."
My guess is OS X.
MORTAR COMBAT!
You're close but you got one thing wrong. PPC601 implements the full POWER instruction set from its day. Every PowerPC processor since then has been missing some functionality (usually instructions) present in POWER.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Don't bastardize a spaceballs quote and try to pass it off as your own!!!
Fundamental misunderstanding re: subsets.
POWER4 is a POWER implementation of the 64-bit PowerPC architecture. This includes QoS/Multi-Core/Hardware-level Virtualization (e.g. ISA extension), etc. All the things that make it an Enterprise-level processor.
PowerPC 970 is a PowerPC implementation of the 64-bit PowerPC architecture. This includes single core/no virtualizations/much faster core timings and most importantly, the ability to run 32-bit PowerPC binaries concurrently with 64-bit PowerPC binaries. All the things that make it a desktop/workstation processor.
That IBM has re-named the POWER architecture PowerPC, does NOT mean that PowerPC==POWER. What it means is that both PowerPC and POWER are PowerPC [Architecture].
A tree is a plant.
A blade of grass is a plant.
Both are implementations of the plant architecture. Now, rename plant, blade of grass.
A tree is not a plant.
There comes a time in every man's life when he must say, "No mother! I do not want any more Jell-O!"
Typo.
A tree is not a blade of grass.
There comes a time in every man's life when he must say, "No mother! I do not want any more Jell-O!"
I suppose we should probably blame IBM for causing some confusion. I won't admit that I'm incorrect, but I'm not sure I have a leg to stand on trying to call you incorrect either. IBM marketing calls the high end stuff "POWER" so there is indeed a distinction, it just isn't officially spedified in a manual.