Cell: Future of Gaming?

[xbdestroya]

Originally posted by: DrMrLordX

Originally posted by: xbdestroya
DrMrLordX: You state that Sony has pushed back the release of Cell-powered products; can you give an example? As far as their five-year development timeline goes, Sony is right on track with Cell's release (so far). If nothing else, THAT's the surprise.

.[/i]

Okay, this is the article of which I was thinking:

http://www.theinquirer.net/?article=9053

It's the Inq, but hey, take it for what it's worth.

LOL, dude - but that article is from 2003!!! Listen, I love the Inquirer, so go read their more recent articles, and you will see Cell is right on track for PS3.

Here's one from today: http://chipzilla.com/?article=20907

It's really about XDR, but explicitly states for use with PS3 and the Cell. If I felt like searchign articles beyond today's, I could easily pull up several hundred for you since that 2003 piece you submitted to prove Cell's going into PS3.

 

[zakee00]

Linkage
havent had time to read it, but i found that link on HardOCP.

 

[vasdrakken]

hummmm so the interconnect between the chips has HOW MUCH BANDWIDTH? Lets assume that it is the same tech AMD and IBM are using so 8 high-speed interconnects right next to each other, so that PCB is going to require twice as many layers as the current AMD/IBM ones something short of 20 layers mobo markers are having trouble with the cost of 9 to 12 layer PCB how much is a 24 layer going to cost? and that totally ignores the interconnects to the GPU, or other onboard chips. Maybe they have found a really effective way to use light interconnects {i.e. like fiber only smaller} or some other tech way out there but one of the biggest things slowing down these massive SMP is that they have to be physically connected. Nvidia's new MCP will be held back by the fact the silicone is going to cost more than any of the individual components since getting silicone to support all of the pathways is going to get more like a large cpu holding smaller modules at this point. Look at a wire diagram of a 8080 or any other early proc they are flat piece of paper, looking at a modern CPU you need a 3D cad program to display them, and the mother boards are already getting there. Look at the wire diagram of a GeForce 6 or ATI X800 series they are multiple layers already. The one thing people are forgetting in a Sony vs. MS is MS can sell all their Xboxes at lose they did it the first time around. The article mentioned he was guessing several times so it is flame bait at best as a fundamental understand how a CPU works seems to missing especially in reference to distributed computing.

 

[Sniper82]

I'll buy a PS3 regardless but it won't be at launch because its gonna have issues like Sony's other 2 systems did at launch. From the sound of things this is gonna be one complicated system so its bound to have issues. Probably not at first but months down the road.

 

[clarkey01]

For some reason my gut tells me Sony/IBM/Toshiba will pull it off. It wont be a 1000x fold increase in performance but It will be a huge step up. I just hope cell is easy to program so developers can get the best out of it. Remember the sega saturn ? was alot more powerful then the PS1 however with 2 cpu's and what was it ? 2 gpu's the thing was quite hard to program. Ever wonder why games like wipeout always looked that bit sharper on the PS1 ?.

I am not going to get involved with all the ?cost? factors that a lot of people are throwing around. If this is just the Sony pr machine working at its best to promote the PS3, there going to extreme and silly lengths. If Cell is not as powerful & complicated as people think, then why do they need to build new fabs and invest billions into it over 5 years ?. Sony could have bought capacity or contracted someone with fabs to make the chip for them. I?m sure they have fabs already but you see where I m going with this. But only time will tell.

 

[zakee00]

Game companies are obviously going to have to get used to coding multi processor/gpu/whatever aware games. It is the future. AMD dualcore, Intel dualcore, Cell, etc....the future IS multi-processors.

 

[DrMrLordX]

Originally posted by: xbdestroya

LOL, dude - but that article is from 2003!!! Listen, I love the Inquirer, so go read their more recent articles, and you will see Cell is right on track for PS3.
.

Whether or not the article was published in 2003 isn't the issue. At issue are the costs involved with Sony readying the Cell for a 2005 release. The Inq guys were convinced that Sony wouldn't be crazy/stupid enough to dump all that extra cash just to keep Cell on schedule.

Did they really drop an extra $3.6 billion to ramp up for full scale production of Cell in 2005? I certainly hope not.

As to the article you linked, all it indicates is that the PS3 will be one of (RAMBUS, inc hopes)many platforms utilizing their new XDR memory. No word on when the chips would go into use in the PS3. Judging by the component shortages Sony experienced at that PS2 launch(memory shortages were a problem, as I recall), it would make sense for Sony to secure large quanities of components for the PS3 prior to launch, if only to avoid the problems of ages past.

 

[xbdestroya]

Originally posted by: DrMrLordX

Originally posted by: xbdestroya

LOL, dude - but that article is from 2003!!! Listen, I love the Inquirer, so go read their more recent articles, and you will see Cell is right on track for PS3.
.

Whether or not the article was published in 2003 isn't the issue. At issue are the costs involved with Sony readying the Cell for a 2005 release. The Inq guys were convinced that Sony wouldn't be crazy/stupid enough to dump all that extra cash just to keep Cell on schedule.

Did they really drop an extra $3.6 billion to ramp up for full scale production of Cell in 2005? I certainly hope not.

As to the article you linked, all it indicates is that the PS3 will be one of (RAMBUS, inc hopes)many platforms utilizing their new XDR memory. No word on when the chips would go into use in the PS3. Judging by the component shortages Sony experienced at that PS2 launch(memory shortages were a problem, as I recall), it would make sense for Sony to secure large quanities of components for the PS3 prior to launch, if only to avoid the problems of ages past.

The Cell is already sampling - it's already being put in prototype workstations and put through it's paces. It's going to be publicly discussed/revealed Feb 6th through 10th at the International Solid-State Circuits Conference. The PS3 is launching in 2006, not 2005. If you still don't believe that the Cell's going to go into PS3, well, I'll just have to dig up more stuff for you.

But if you feel a quick ramp isn't worth the expense to Sony, you're wrong. PS3 has to be successful this generation. The console business is too much a traditional linchpin of Sony's profitability to be cut slack just when MS is starting to turn up the heat. Cell may rule the world or it may suck, but whatever the case, expect it in PS3.

 

[TekDemon]

Originally posted by: mwmorph1. THe screenshots are faked. no way they have techdeoms for soeething a year from release
2. not as impressive as the scrnshots? isnt the gscube SUPPOSED to be more powerful? shouldnt they look BETTER?

it's obvious sony is lying it's ass off

Huh? The GSCUBE is based on PS2 tech...it's been around for a while now so it wouldn't make any sense that it's more powerful.

The Screen Shots aren't faked, they're not PS3 Screen Shots, but GSCUBE realtime renders.

 

[DrMrLordX]

Originally posted by: xbdestroya


The Cell is already sampling - it's already being put in prototype workstations and put through it's paces. It's going to be publicly discussed/revealed Feb 6th through 10th at the International Solid-State Circuits Conference. The PS3 is launching in 2006, not 2005. If you still don't believe that the Cell's going to go into PS3, well, I'll just have to dig up more stuff for you.

But if you feel a quick ramp isn't worth the expense to Sony, you're wrong. PS3 has to be successful this generation. The console business is too much a traditional linchpin of Sony's profitability to be cut slack just when MS is starting to turn up the heat. Cell may rule the world or it may suck, but whatever the case, expect it in PS3.

By now, it's pretty obvious Sony is serious about releasing Cell. I don't doubt that. What I do wonder about is the capabity and cost of the chip itself. Honestly, Sony could have used other CPUs in the PS3 to make it successful. Given the fact that Cell will be used in a wide variety of products(not just the PS3), it would seem to me to be wiser to put off release until 2007, when it wouldn't be so damn expensive for them to do it. $3.6 billion is a lot of money, even for Sony.

Either way, they may have the next z80 on their hands. Or they may not.

 

[anthrax]

Originally posted by: DrMrLordX

Originally posted by: xbdestroya


The Cell is already sampling - it's already being put in prototype workstations and put through it's paces. It's going to be publicly discussed/revealed Feb 6th through 10th at the International Solid-State Circuits Conference. The PS3 is launching in 2006, not 2005. If you still don't believe that the Cell's going to go into PS3, well, I'll just have to dig up more stuff for you.

But if you feel a quick ramp isn't worth the expense to Sony, you're wrong. PS3 has to be successful this generation. The console business is too much a traditional linchpin of Sony's profitability to be cut slack just when MS is starting to turn up the heat. Cell may rule the world or it may suck, but whatever the case, expect it in PS3.

By now, it's pretty obvious Sony is serious about releasing Cell. I don't doubt that. What I do wonder about is the capabity and cost of the chip itself. Honestly, Sony could have used other CPUs in the PS3 to make it successful. Given the fact that Cell will be used in a wide variety of products(not just the PS3), it would seem to me to be wiser to put off release until 2007, when it wouldn't be so damn expensive for them to do it. $3.6 billion is a lot of money, even for Sony.

Either way, they may have the next z80 on their hands. Or they may not.

Cell technology for the PS3 at least is mostly SONY HYPE HYPE and more HYPE.

For multiplayer gaming applications (i.e. multiuser multimedia interactive entertainment applications) you will want minimal latency. (i.e. minimum amount of time between user input and the display of the results on the players screen)..

Now, you notice that cell architecture depends that Network to send the (program + data) to another cell....

Within a single device, it might be practical to implement it with a certain dedicated bus of some type.... over a highspeed bus like hypertransport of Front Side Bus or something custom designed...

But lets say you want to play a multiplayer game between two PS3 users across a city... The network is still your bog standard IP network with high latency (compared to the speed of a typical front side bus found with in a computer.)..............

In other words, CELL technology has very little benifit to gamming applicaitons...

The main benifit of "CELL technology" is that programs will be much easier to scale over large number of processors... but that depends on the type of applicaitons...

For applications with simple instructions and a large data set (i.e. scientific applications) it will work....

 

[UbiSunt]

-For further reading here is a list of Blanchford's favorite books (from his own site) on UFO's:

http://www.blachford.info/other/stuff.html

UFO

Open Skies Closed Minds - Nick Pope
Nick Pope was the UK Ministry of Defences official UFO investigator for three years and didn't think much about the subject until he was given the Job. Most UFO reports turn out to be planes or planets or all sorts of other things but there are some (5-10%) which are a real mystery. Nick Pope investigated these and found that there were some questions to be asked and indeed concluded that we were indeed being visited by someone or something from above. This book is fascinating and got me and probably many others interested in the subject (I've been reading UFO Magazine ever since). It turns out there is a lot more to UFOs than the odd light in the sky and he is not the only official who is convinced that we are being visited from above. There are admittidly a lot of nuts in the UFO field but Nick Pope is not one of them, this book lets you know what's really going on.

The Uninvited - Nick Pope
This is Nick Pope's second book which this time deals with the subject of so called Alien abductions. A better written book than his first and a well researched book which goes into some detail about contact with what he calls "non-human" intelligence's. This book covers all sorts of theories and serves as a brilliant introduction to the abduction subject as a whole. He presents a guide to the general experiences said to happen then relates a number of new rather strange cases which he has investigated. Overall a fascinating and well written book which leads you into the bizarre world of alien abductions and all that goes with them. The ramifications of some of the subjects raised are at the very least shocking. Whatever your opinion of these so called abductions you can be sure of one thing - something truly bizarre is and has been going on for a very long time.

Out There - Howard Blum
Another UFO book but this time from the other side of the pond. This guy is an award winning author and got into this almost by mistake but did a great deal of investigation to write a very interesting book. It tells the story of a military group set up to investigate the subject and some of what they investigated. It does drag on a bit but it gives very good background information about how the US authorities have been investigating UFOs despite all their denials (including very recent ones).

-This guy is a self-professed dabbler in a lot of disciplines that he is not involved in. He might be highly intelligent but he also seems highly neurotic. I'm going to wait until Anandtech posts some actual benchmarks before I make any further conclusions about Cell technology.

-Perhaps, aliens created Cell technology and licensed it to Sony

 

[DrMrLordX]

Originally posted by: UbiSunt
-Perhaps, aliens created Cell technology and licensed it to Sony

If David Duchovny shows up at the International Solid State Circuits Conference . . . hoo boy!

The truth is out there.

 

[VirtualLarry]

Originally posted by: anthrax
Cell technology for the PS3 at least is mostly SONY HYPE HYPE and more HYPE.

For multiplayer gaming applications (i.e. multiuser multimedia interactive entertainment applications) you will want minimal latency. (i.e. minimum amount of time between user input and the display of the results on the players screen)..

Now, you notice that cell architecture depends that Network to send the (program + data) to another cell....

Yes, true... BUT... you need to consider who is partnering with Sony and Toshiba on this ... and what very same company is promoting their "grid computing" technology.

Perhaps the real benefit isn't going to be for PS3 owners/users, per-se, but rather IBM.. if there is an installed-base of 10mil of these 4-Cell units, most connected to always-on network connections... that's a lot of CPUs to potentially be able to run distributed-computing apps on, and for IBM to broker that CPU time to grid-computing clients of theirs.

It should be interesting to read the user-agreement that comes with the PS3, or as part of the TOS for accessing online games with it - if Sony/IBM reserves the right to freely use your spare CPU cycles, should it be left connected and powered-on. Supposedly, the PS3 won't even have a physical off switch, it will be purely soft-off only, like most televisions and VCRs.

Originally posted by: anthrax
In other words, CELL technology has very little benifit to gamming applicaitons...

The main benifit of "CELL technology" is that programs will be much easier to scale over large number of processors... but that depends on the type of applicaitons...

For applications with simple instructions and a large data set (i.e. scientific applications) it will work....

I totally agree, and the customers being sold the gaming angle of Cell, are indeed mostly being sold hype. But the distributed-computing potential of these units, once produced in the millions, should be quite substantial. I wonder if end-users will be compensated? (More than likely it will be a tradeoff, in order to access Sony online services, you have to agree to "sell" your spare CPU time. In fact, for some uses, your unit wouldn't even have to be online all of the time, it could download work units and upload results.)

 

[VirtualLarry]

Originally posted by: xbdestroya
I am on several forums that debate the nature of the Cell chip and I have lifted this post from one of them. I'm not sure if someof you will consider it too technical, but read it and consider the implications for die-size. This guy does a lot better job of explaining than I can:
Anyway, I figure the whole idea of CELL is that the APUs would be very simple in nature to the effect that the APU by itself is not a great performer, but it's cheap to make and very low transistor budget, and a whole collection of APUs makes for some high overall throughput. That suggests to me that these APUs would not be loaded with all nature of extra performance boosters like out-of-order-execution, branch prediction, SMT, multiple issue, etc. They're essentially very basic straight-forward single-issue pipelines. That in turn, makes the pipelines and each pipeline stage very simple. So with that kind of simplicity, high clocks are very easy to achieve. BTW, I don't know how much stock I put in the idea that the APUs would be VLIW in the effect of being able to issue SIMD floating point and integer instructions in parallel. I actually find it easier to believe that you can only issue instructions on separate clock cycles regardless of type.

That actually makes a lot of sense, a "throughput-oriented" architecture, simple in nature, but made up for in numbers. I was reading a paper on the overall "Google cluster" architecture, and the overall design philosophy that was outlined was very similar. Each commodity machine used in the cluster was relatively simple, and didn't have the fancy (and expensive) features that high-end servers did, like redundancy, etc., because the overall entire architecture took that all into account.

This may also translate into a mfg advantage as well - much like current high-end GPUs, if one of the processing units proves to be defective during testing, it can be disabled/mapped out by the mfg testing process. There could be otherwise-identical Cell chips rolling off the lines, with different numbers of enabled/working pipelines, and some of the less-powerful Cell chips might end up being sold/used in things like televisions rather than the PS3.

 

[VirtualLarry]

Originally posted by: xbdestroya
It's going to be scale scale scale. There's going to be more fabs producing Cell chips than AMD presently has capacity period. PS3 alone over a five-year period is expected to account for 100 million units; and if it really is four per system, whcih I doubt it would be, 400 million Cell chips right there. To say nothing of the other things it will be going into.

Wow, that's... crazy, when you think about it, that sort of level of scale. Even today, with x86 being pervasive at least in the world of PCs - but your phone doesn't have an x86, nor your TV, nor your microwave. (more likely a Hitachi S8 or PPC or Coldfire, but I digress).. in terms of true economies of scale, Cell should be a juggernaut. Here's a neat question - will Cell have onboard or on-chip wireless capability? Will your "Cell" phone (ha!) have the ability to talk to your TV and your microwave, and crunch data, when you're not home? What about talking to your neighbor's Cell-enabled devices too? Btw, I just watched the entire Bubblegum Crisis Tokyo 2040 series quite recently... if you know what the ending to that is like, and you compare it to Cell, if one added inbuilt wireless networking... hmm.

Originally posted by: xbdestroya
I think there's frequently a misconception that just because something is new and awesome it's expensive. That's not the case. It doesn't cost Intel hardly any more to make a Prescott than a Celeron, but they price it differently because that's the market and that's the industry they are in. Sony is not in this to sell these processors to others, so they will use these chips as a means to an end, not the end itself.

I'm glad that someone understands large-scale contract pricing on things. Even though it is new tech, each Cell chip doesn't necesarily have to cost as much as an X800XTPE for example, the high costs of the initial R&D are amortized over the entire production and the price/unit is amortized over the length of the contract. Spot demand doesn't factor in in the same way that it does for individual units on the open market.

 

[Avalon]

Originally posted by: VirtualLarry
Btw, I just watched the entire Bubblegum Crisis Tokyo 2040 series quite recently... if you know what the ending to that is like, and you compare it to Cell, if one added inbuilt wireless networking... hmm.

 

[R3MF]

Originally posted by: R3MF

Originally posted by: UzairH
The article makes no mention of how many transistors each APU and the associated memory logic is going to have. Even taken conservatively, 8 APUs plus 1 PU plus all the other switches/logic seems like it would come to over 500 million trasistors. Now how the hell will the power consumption be kept low with so many transistors working at 4.6GHz?

nice assumption, but does it have any validity at all?

it has no cache to speak off, certainly not the 2MB that x86 processors come equipped with, which takes up about 75% of the die space.

how many transistors does an ARM CPU have? 25 million, 30 million? that is probably a reasonable top-end guess for what an APU might take up..........

here is my wild assumptions:
1x PPC core = 50million
8x APU = 200million
logic+IO = 50million

there you go, 300million trannies all in, not much bigger than a NV40 GPU.

here is hoping Ubuntu make a Cell a target architecture.

Originally posted by: R3MF
i personally think that 300million is an overestimation, with 320million being absolutely the top whack, with 250million or less being more likely.

232m confirmed.

vindicated.

 

[clarkey01]

I hear this cell will BEAST a celeron 266 Mhz.....Seshhs...


Wasnt it supposed to be talked about yesterday ? by IBM/Sony ?

 

[Acanthus]

Im more interested in the NVIDIA core than the Cell chip to be honest

It is very likely an NV50 derivative will be what we see.

 

[xbdestroya]

Originally posted by: clarkey01
I hear this cell will BEAST a celeron 266 Mhz.....Seshhs...


Wasnt it supposed to be talked about yesterday ? by IBM/Sony ?

Yeah, it WAS talked about - and it's being talked about all week.

For a simple link, go here - http://anandtech.com/news/shownews.aspx?i=23781

 

[Eug]

Many people are suggesting that the PPE core of Cell is very similar to one of Xbox 2 Xenon's three CPU cores.

Maybe it should be "PPE: Future of Gaming?"

 

[clarkey01]

Erm, for some reason I think this wont be all hype, I too tired to go into all the detail but I just have a feeling it'll knock the pants off the Xbox 2( xbox next) and the Nintendo Revo.

Funny how IBM is making chips for all 3 of them, there was talk of cell ending up in the xbox 2 aswell. Sorry Iv been out of the circle for a while on the cell.

 

[91TTZ]

Originally posted by: zakee00
Give an example of how Sony has lied about their CPU's.

When the PS2 was about to come out, Sony claimed that the Emotion Engine was powerful enough to render Toy Story in realtime.

That was a complete lie.

 

[91TTZ]

Hey guys, remember reading any of Anand's reviews on CPU's? On the graphics tests, you often see that all the CPU's get the same framerates if the video card isn't fast enough. It becomes a bottleneck and you'll be fill rate limited.

I think all the talk about the Cell is hype, but let's just pretend that it's not for a moment. Even if it's 200 times faster than any other chip, it will eventually end up being limited by the graphics chipset. Try running an Athlon64 4000+ with a GeForce2 MX and see what kind of gaming performance you get. You'd be better off having the CPU render the frames.

That is, unless, Sony claims that the GPU in the PS3 is a super awesome revolutionary GPU of epic proportions, like they do to all their other stuff.