Does Cell Have Any Other Advantages Over XCPU Other Than FLOPS?

[Shifty Geezer]

Sorry but this all compeltely far fetched shifty, this stuff just isn't going to happen. I'm looking for realistic scenarios here, I specifically asked NOT to blue-sky, but to give realworld achievable examples.

Did you read the entirety of my post? As I said, the ideas that'll be supported by Cell are things not thought of yet, in the same way modern processors are doing things that older processors would never have done. If we had stopped with 486s which were perfectly adequate for the tasks of the time, we wouldn't be talking about media functionality at all.

 

[nelg]

IMO, custom ASICs will always be the rule for consumer devices. The shear volume means that development cost are amortized to the point of being trivial. They also have better size, thermal and power consumption characteristics.

 

[Vince]

Why use Cell for decoding H.264, when you can just use a custom ASIC for a fraction of the cost and a fraction of the power consumption?

Your argument is against the trend in CE equiptment which is moving additional functionality onboard. Sony's VME is a good example, except that the VME is designed where power consumptions actually matters, unlike in a CE product -- especially a high-end one.

Can that ASIC also upconvert your DVD input to HD? What image enhancements can it do? Can it simultaneously encode/decode any audio standard in the CE market? Can it process a visual/augmented-reality based GUI/menu system? How many HD streams can it decode? Will it allow for Video-over-IP or VoIP? Can it be used to compress and store several HD streams to a HDD, downconvert and stream one over IP (say, to a PSP or Cell-phone). What about allowing for basic broadband connectivity processing (Connect ready?!?), single port plug-and-play compatability with any other Sony CE part equipt with Cell....

How many discrete ASICs are we upto now?

Interestingly this is exactly what the issue with the Nuon buiness model was.

I don't exactly think the situations are analogous. Sony Group itself, today and increasingly looking forward, is it's own consumer of semiconductors. They create their own demand and are doing it, as the most recent IC Insights rankings demonstrate, at an increasing rate.

Just look at the fact that of the top 10 semiconductor companies, only 4 beat the average market growth rate: Sony, Intel, Samsung, and Toshiba. 3 of them are CE companies (and Intel is Intel); the trend is quite overt.

 

[swaaye]

Just think about the little $50-60 Philips DVP642 Divx set top player that is everywhere now. It pulls about 11W top from the outlet (I measured) yet can play something that takes about a P3-500 machine to play. A P3-500 machine will pull around 80W from the outlet.

The chip inside this unit doesn't even need a heatsink.

Ya, ASICs are useful.

 

[scooby_dooby]

Did you read the entirety of my post? As I said, the ideas that'll be supported by Cell are things not thought of yet, in the same way modern processors are doing things that older processors would never have done. If we had stopped with 486s which were perfectly adequate for the tasks of the time, we wouldn't be talking about media functionality at all.

I understand that in terms of gaming applications.

In terms of multimedia applications, which is what the discussion was about, the console is launching in months, and if these multimedia applications haven't even been thought of then obviously they won't get implemented.

What does CELL honestly bring to the table as a multimedia processor? Keeping in mind bandwidth limitations over the next 5 years?

It's great if you wanted to encode video all day, but the consumer(gamers) are not interested in that. WHat does it bring to the table that makes it so much better for the average consumer? What will it tangibly be able to do that will set it apart from the competition?

 

[aaaaa00]

Can that ASIC also upconvert your DVD input to HD? What image enhancements can it do? Can it simultaneously encode/decode any audio standard in the CE market?

It's got enough acceleration hardware to decode all of the codecs that matter (the mandatory ones for BD or HD DVD) with the number of streams that matter (1, or 2 if PiP is supported), and enough image processing hardware to do the deinterlacing and image enhancement that is necessary.

Can it process a visual/augmented-reality based GUI/menu system? How many HD streams can it decode?

Who cares? You don't need either of those for a $50 next-generation DVD player.

Will it allow for Video-over-IP or VoIP?

Sure. Neither takes gobs of processsing power. You'll note there's a 200 mhz ARM core in that ASIC, which when combined with the graphics hardware, should supply enough horsepower for any reasonable movie delivery interaction layer.

Can it be used to compress and store several HD streams to a HDD, downconvert and stream one over IP (say, to a PSP or Cell-phone).

J6P doesn't give two craps about this stuff. They want their $50 DVD player, only in HD. In fact most people hate dancing menus and interactivity features, they just want to stick the movie in and have it play.

Cheap, reliable, and fulfills the basic requirement of playing good looking HD movies, that's all that's necessary.

What about allowing for basic broadband connectivity processing (Connect ready?!?),

Connect? Who cares.

single port plug-and-play compatability with any other Sony CE part equipt with Cell....

This only matters if you care about compatibility with Sony CE equipment. Most people won't have all Sony equipment anyway.

CELL is good for a console where all this processing power is actually demanded. The fact you can run a software H.264 codec on it really well is a nice bonus for PS3, but that doesn't make CELL a good replacement for the ASICs in cost sensitive CE products.

Most CE devices do one thing reliably, simply, and cheaply -- custom ASICs tend to be the most reliable, simple, and cheap way to do that, which is why they vastly outnumber general purpose CPUs there and will continue to do so IMHO.

There is zip-nada-nothing Sony will be able to do that will bring PS3 down to anywhere near the bill-of-goods for a next-generation DVD player built around custom ASICs in a similar time frame -- you can buy entire DVD players for $25 retail these days (barely more than the cost of a bare drive), and similar cost scaling is pretty much guaranteed for HD DVD/BD ASICs.

 

[Fox5]

Why not seing 2 chanells at the same time while with sorrund sound in the CPU while the imge is treated to get even better, recording one of them (to the HDD) and use voice control instead of a remote, there is plenty of uses in a CELL beside gaming (and no, this will/would not be a total chaos, what we do everyday in a PC at the same time like Foruns, listen to music, encoding, download... could also look a chaos 15 years ago).

Plus it is supossed to came a HD camera with every PS3.

People in theese cases should think it the future not in the present.

You think Sony is really going to be developing all that software for PS3? Or will allow others to even? Software is king, regardless of what the hardware can do.

[q]A P3-500 machine will pull around 80W from the outlet.[/q]

There are some low voltage centrino systems that take less than 20 watts to power the system (at around 1ghz I guess) and screen.

 

[Shifty Geezer]

It's great if you wanted to encode video all day, but the consumer(gamers) are not interested in that. WHat does it bring to the table that makes it so much better for the average consumer? What will it tangibly be able to do that will set it apart from the competition?

That seems to me like the PC argument though. 99% of the time a PC user's resources are 95% unused. Photoediting is a good example. Most functions are fairly instantaneous, but then you get these fancy plugin effects that can take a long time (tens of seconds). One reason to upgrade from a P4 2GHz to P4 3GHz is so that in that 1% of the time you use the system resources it runs faster. I don't know that outside of games PCs can do anything now that they couldn't with the tech they had 2 years ago. My AthlonXP 2500 can do everything my friends' Athlon64 can do, just some things not as quick. But this upgrade pathway for very limited gains seems to have pushed processor advances and consumers have shown that they will pay for faster, even when the benefits are actually very small.

As for tangible features Cell brings to CE, that discussion was covered in another epic thread where I gave some points such as development and reuse of software routines between industries. As you were around then I'm pretty sure you partook in that discussion so there's no point revisiting it; least not in this thread.

 

[Vince]

It's got enough acceleration hardware to decode all of the codecs that matter (the mandatory ones for BD or HD DVD) with the number of streams that matter (1, or 2 if PiP is supported), and enough image processing hardware to do the deinterlacing and image enhancement that is necessary.

That was yesterday, what can you give me today?

]Who cares? You don't need either of those for a $50 next-generation DVD player.

Why do I get the feeling this is going to be a typical answer? Secondly, $50? We're talking about next generation products: Blu-Ray, HD-DVD, 1080i/p broadcast quality. Where did the $50, POS chinese knock-off player enter into this argument?

Cheap, reliable, and fulfills the basic requirement of playing good looking HD movies, that's all that's necessary.

Connect? Who cares.

Figures! Miss Cleo move over... Ohh, and the [cheaper] > [quality] argument really worked in the recent case of HD-DVD verse Blu-Ray, huh?

The facts are simple. For CE makers like Sony and Samsung who are targeting the lucritive buyers who actually have disposable money to spend (eg. not your shit-ass knock-off players from China), it's clear that people are demanding more functionality, more features and options, and a greater ease of use; and that this needs to be combined with the realization that processign requirements are increasing as HDTV is on the verge of exploding in adoption starting this Christmas and beyond.

Discrete ASICs are not going to work, they will move into the same niche as the concept of a DSP array for every task. Logic is inexpensive and as capacity increases, computation is decreasing in cost to the point where someone like Sony can replace their discrete DSPs and ASICs with a Cell-esque IC which does all the functions of the DSPs, have added value (eg. look at PSP) and can be updated to support new standards, connect to their network fabric and offer all the little bells and whistles that people who actually spend money demand.

If you want to argue over the realm of the bottom-feeder, then perhaps a new forum would work for you; but this is a technology-orientated forum. We're discussing the technologically advanced solutions, not the shitty equiptment you're talking about.

 

[Shifty Geezer]

You think Sony is really going to be developing all that software for PS3? Or will allow others to even? Software is king, regardless of what the hardware can do.

KK talked of having things like image and video editing software for PS3. As for where it comes from, I'm not overly impressed with Sony's software efforts to date and question their capacity to provide anything worthwhile. But the open attitude towards Cell thus far and possibility of Linux strongly suggests Sony are looking for 3rd party software development of applications. Certainly if I had the wherewithall I'd grab a copy of the Cell SDK and set about writing an accoustic-modelling software synthesizer or image manipulator for Cell+Linux. Unlike the PC space where every product has to compete with a dozen rivals ranging from expensive pro tools to shareware and freeware, PS3's Linux software situation will be a first-come, first-served market where the early entrants can make a name for themselves as the only software options. Anyone who can provide tools that effectively harness Cell's 'media' performance in common media applications is going to find custom.

 

[mckmas8808]

Thank you Vince I got tired of hearing about dirt cheap $50 DVD players.

 

[aaaaa00]

Why do I get the feeling this is going to be a typical answer? Secondly, $50? We're talking about next generation products: Blu-Ray, HD-DVD, 1080i/p broadcast quality. Where did the $50, POS chinese knock-off player enter into this argument?

The fundemental truth is that DVD didn't hit mainstream until vast quantities of cheap players hit the market. I don't anticipate either of the new formats will reach critical mass until the same thing happens.

Most CE devices exist in markets that are extremely price-elastic because they are non-essential luxury goods. That is why CE manufacturers are always extremely cost-sensitive with designs that they intend to hit mainstream. There is a small market of early adopters and tech-heads, but aiming for them is only a guarantee you will always be a niche product -- like Laserdisc, DVD-A, and SACD.

 

[scooby_dooby]

can be updated to support new standards, connect to their network fabric and offer all the little bells and whistles that people who actually spend money demand.

Where is this demand? In a tiny niche of technology enthusiasts? That's not a wide enough audience to justify the cost.

I don't see this demand you speak of, people want affordable reliable solutions with improved image quality.

Bells and whistles are just that, frivolous, there is no reason to believe there is a huge demand for any of these 'extras' (of which I still haven't heard an decent concrete examples) If there's no signifigant demand I don't know why you think Sony or Toshiba is going to waste money of cell processors in their CE devices.

btw - what happened to those Toshiba CELL tv's slated for 06? Gone bye-bye?

 

[scificube]

If the bells and whistles come at no extra cost? They will in the PS3 or are we not talking bout the PS3 in this instance?

In CE products it still comes free with reject Cells and other CBEA implementations targeted directly towards them. A 1:1 CBEA seems just fine in most cases no?

 

[expletive]

The facts are simple. For CE makers like Sony and Samsung who are targeting the lucritive buyers who actually have disposable money to spend (eg. not your shit-ass knock-off players from China), it's clear that people are demanding more functionality, more features and options, and a greater ease of use; and that this needs to be combined with the realization that processign requirements are increasing as HDTV is on the verge of exploding in adoption starting this Christmas and beyond.

Where is it clear that people are demanding more functionality? What exactly are they demanding from their TVs other than higher resolution? Memory card slots?

Most of my friends dont even notice/care about crap video deinterlacing (seen by horrible jaggies) when its right in front of their face on their HDTVs, much less anything above and beyond the most basic functionality.

Of all the things you mention here, features, options, ease of use- NONE of those are only made possible by the use of a cell processor. These things are products of good design first and foremost, and technology second.

ive got a chip in my cable box that can scale/deinterlace/PIP all in HD. Maybe its not the greatest but i guarentee you that 99% of the consumers couldnt care less about processing an order of magnitude better than what this provides. Ive got a $1500 dedicated scaler in my home theater that i use on a 110" screen. Problem is, i may be the only person in the room that realizes what improvements it it adds to the image, im sure the rest of my guests would be jsut as impressed with a direct feed from the cable box.

Discrete ASICs are not going to work, they will move into the same niche as the concept of a DSP array for every task. Logic is inexpensive and as capacity increases, computation is decreasing in cost to the point where someone like Sony can replace their discrete DSPs and ASICs with a Cell-esque IC which does all the functions of the DSPs, have added value (eg. look at PSP) and can be updated to support new standards, connect to their network fabric and offer all the little bells and whistles that people who actually spend money demand.

Discrete ASICS have worked, will work, and are still working. Dedicated chips will always be able to do the same required functionality for a lower transistor count, power, and cost. If it makes more sense to Sony to use Cells in everything they make, thats peculiar to Sony becuase of their investment in it - but does not guarentee any competitive advantage whatsoever. Just because its good for Sony doesnt necessarily make it good for the consumer.

Yeah Sony would sure like to sell 100,000 $25k qualias every year but the fact of the matter is that the margin may be there but the market isnt. Are you saying Sony's strategy is to sell nothing but $3000 TVs with cell processors? That kind of thinking got them into the financial trouble in the first place.

If you want to argue over the realm of the bottom-feeder, then perhaps a new forum would work for you; but this is a technology-orientated forum. We're discussing the technologically advanced solutions, not the shitty equiptment you're talking about.

I think he's talking about reality and, imo, seperating it what the tech savvy imagine will happen in the market just becuase its "technologically ideal". But hey, he doesnt need me to fight his battles.

 

[nelg]

There is a saying in business.........

Sell to the masses and live with the classes.
Sell to the classes and live with the masses.

 

[seismologist]

You know, Companies dont spend billions of dollars on new technology without having applications in mind.
It seems like it would be more productive to think how how they intend to use Cell instead of trying to fight it. Since It's a losing battle.

There was a quote from some Sony exec on why a scalable architecture like Cell is needed going forward as opposed to conventional Asic. Lets see if I can find it...

 

[expletive]

You know, Companies dont spend billions of dollars on new technology without having applications in mind.
It seems like it would be more productive to think how how they intend to use Cell instead of trying to fight it. Since It's a losing battle.

There was a quote from some Sony exec on why a scalable architecture like Cell is needed going forward as opposed to conventional Asic. Lets see if I can find it...

I dont think anyone is saying that Sony doesnt know how theyre going to use it, only that theyre not sure if it will provide a tangible benefit to the mass market consumer...other than FLOPS.

A Sony exec saying the cell is needed going forward is like bill gates saying vista is needed for PCs to move forward. Don't waste your time looking.

 

[nelg]

You know, Companies dont spend billions of dollars on new technology without having applications in mind.
It seems like it would be more productive to think how how they intend to use Cell instead of trying to fight it. Since It's a losing battle.

There was a quote from some Sony exec on why a scalable architecture like Cell is needed going forward as opposed to conventional Asic. Lets see if I can find it...

That does not guarantee success.

 

[seismologist]

That does not guarantee success.

Well at least they had an application in mind. It's not like somebody at Sony one day decided for no reason that they need to design a CPU with more Flops...